Citation
Current-accelerated channel hot-carrier stress in silicon metal-oxide-semiconductor transistors

Material Information

Title:
Current-accelerated channel hot-carrier stress in silicon metal-oxide-semiconductor transistors
Creator:
Han, Kim-Kwong Michael, 1966-
Publication Date:
Language:
English
Physical Description:
viii, 164 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Charge carriers ( jstor )
Drains ( jstor )
Electric current ( jstor )
Electric potential ( jstor )
Electrons ( jstor )
Hot electrons ( jstor )
Kinetic energy ( jstor )
Oxides ( jstor )
Silicon ( jstor )
Transistors ( jstor )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1997.
Bibliography:
Includes bibliographical references (leaves 151-163).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Kim-Kwong Michael Han.

Record Information

Source Institution:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
38863551 ( OCLC )
ocm38863551

Downloads

This item has the following downloads:


Full Text














CURRENT-ACCELERATED
CHANNEL HOT-CARRIER STRESS IN
SILICON METAL-OXIDE-SEMICONDUCTOR TRANSISTORS









BY

KIM-KWONG MICHAEL HAN










A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1997
















ACKNOWLEDGEMENTS



I wish to express my gratitude to Professor Chih-Tang Sah for all his wisdom, guidance, patience and time throughout the course of my graduate research at the University of Florida. I would also like to express my thanks to Professors Arnost Neugroschel, Toshikazu Nishida, Peter Zory and Ulrich Kurzweg for serving on my supervisory committee.

I wish to thank my past and present colleagues, Dr. Scott Thompson, Dr. Yi Lu, Dr. Jack Kavalieros, Dr. Michael Carroll and Steven Walstra, for helpful discussions. Special thanks to Steve Walstra for his friendship and the many interesting and insightful discussions and debates. Assistance by Derek Martin and Jin Cai is also appreciated. Financial support from Semiconductor Research Corporation is gratefully acknowledged.

I am indebted to my family for their persistent caring and support throughout this endeavor.















11
















TABLE OF CONTENTS

page

ACKNOW LEDGEM ENTS.......... ............................................................................ ii

A B ST R A C T ....................................... ............................................................ ........ vi

CHAPTERS

1 INTRODUCTION............................ ......................... 1

2 STRESS-AND-MEASURE (SAM) METHODOLOGY....................... 7

2.1 Introduction.................................................................................... 7
2.2 SA M Setup.................................................................................. 7
2.3 Electrical Measurement............................................................ 9
2.4 Substrate and Channel Hot Electron Stress............................... 15

3 DIRECT-CURRENT CURRENT-VOLTAGE (DCIV)
CHARACTERIZATION TECHNIQUE.............................. ......... 26

3.1 Introduction.............................................................................. 26
3.2 Measurement Configurations........................... ............. 27
3.3 C om ponents of IB........................................................................... 29
3.4 Theory of Surface Recombination Rate................................. 33
3.5 Relationship between xys and VGB under Non-Equilibrium
Condition.................................................... 40
3.5 Sum m ary.................................................................................. 46

4 SEPARATION OF INTERFACE AND OXIDE TRAPS USING THE
DCIV TECHNIQUE.................................................. 48

4.1 Introduction.............................................................................. 48
4.2 Experim ent............................................................................. 49
4.3 Results and Discussions.......................... ..................... 55





iii









5 PROFILING OF INTERFACE TRAPS BY THE DCIV METHOD ....... 61

5.1 Introduction................................... ............................................ 6 1
5.2 T heory........................................ ............................................... 62
5.3 Experiment and Results.................................. ............. 67

6 LINEAR REDUCTION OF DRAIN CURRENT WITH INCREASING
INTERFACE RECOMBINATION IN NMOS TRANSISTORS
STRESSED BY CHANNEL HOT ELECTRONS............................. 75

6.1 Introduction............................................................................. 75
6.2 Experim ents.......................................................................... 75
6.3 Results and Discussion.......................................... 76


7 INTERFACE TRAPS GENERATION MODEL................................ 83

7.1 Introduction............................................................................. 83
7.2 T heory......................................... .............................................. 84
7.3 Experiments and Results.................................. ............. 90

8 PHYSICS-BASED TIME-TO-FAILURE EXTRAPOLATION
ALGORITHM USING CURRENT-ACCELERATED CHANNELHOT-CARRIER STRESS................................................................... 95

8.1 Introduction........................................................................... 95
8.2 Methodology of Current-Accelerated Channel-Hot-Carrier
Stress........................................................... 96
8.3 Results and Discussions ........................................ 100

9 HOT HOLE INJECTION INTO SiO2 IN N-CHANNEL MOS
TRANSISTOR DURING CHANNEL-HOT-ELECTRON STRESS....... 117

9.1 Introduction............................. 117
9.2 Four Fundamental Interband Hot Hole Generation Pathways..... 119 9.3 Experim ents........................................................................... 122
9.4 Results and Discussions.............................................. 123

10 REDUCTION OF INTERFACE TRAPS IN P-CHANNEL MOS
TRANSISTOR DURING CHANNEL-HOT-HOLE STRESS ............... 138

10.1 Introduction............................. 138
10.2 Experimental Results .............................. 138
10.3 Kinetic M odel.................... ................. 143
10 .4 Sum m ary...................................................... ............................ 147


iv









11 SUMMARIES AND CONCLUSIONS........................ 148

R EFER E N C ES................................................................................ ... ............... ..... 151

BIOGRAPHICAL SKETCH.................................. 164


















































V1
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CURRENT-ACCELERATED
CHANNEL HOT-CARRIER STRESS IN SILICON METAL-OXIDE-SEMICONDUCTOR TRANSISTORS by

Kim-Kwong Michael Han

December 1997



Chairman: Prof. Chih-Tang Sah Major Department: Electrical and Computer Engineering

Transistor performance degrades during operation due to charging and generation of oxide and interface traps by hot carriers. A novel direct-current current-voltage (DCIV) technique is employed in this thesis to investigate the electrical degradation kinetics of metal-oxide-semiconductor transistors (MOSTs) under hot-carrier stress. This technique measures the interfacial recombination current Ig (base or body current) as a function of the gate voltage to monitor the generation rate of interface traps and charging rate of oxide traps. Two DCIV applications are demonstrated: (1) separation of stress-generated oxide and interface traps in submicron MOSTs, and (2) profiling the spatial distribution of of the stress-generated interface traps. It is also shown that stressgenerated surface recombination current peak is proportional to the traditionally



vi









monitored transistor degradation parameters: drain-saturation current, gate threshold voltage, and subthreshold voltage swing (AIDsat, AVGT and AS).

A physics-based Time-to-Failure extraction algorithm is developed using the new current-accelerated channel-hot-carrier stress (CACHC) method. Current-acceleration of the transistors' degradation rate is obtained by forward-biasing a p/n junction to increase the channel hot carrier current. The applied stress voltages can be independently set to control the kinetic energy of the channel hot carriers. Stress-time reductions by about 1-2 orders of magnitude are demonstrated. Correlation with Sah's bond-breaking interface generation model showed that interfacial silicon-hydrogen bonds are broken by hot holes in n-channel MOST and by hot electrons in p-channel MOST.

Positive charging of oxide trap by barrier-surmounting hot hole during channel hot electron stress is investigated using the DCIV technique. Two pathways are experimentally observed: (1) hot-electron/hot-hole interband impact pathways in highvoltage long-channel nMOST with electron kinetic energy threshold of 5.12 eV, and (2) hot-electron/thermal-hole Auger recombination pathway in low-voltage short-channel nMOST with electron kinetic energy threshold of 3.13 eV.

Interface trap generation in p-channel MOST during channel hot hole stress is investigated. Reduction of interface trap density is observed for the first time. Hydrogen released from the hydrogenated boron in the p+drain region by the channel hot hole is the proposed mechanism. The released hydrogen diffuses from the p+drain






vii









into the channel region and passivates the fabrication-residual interface traps over the channel and source interface.


















































viii















CHAPTER 1
INTRODUCTION

Recent advancement in silicon Very-Large-Scale-Integration (VLSI) and UltraLarge-Scale-Integration (ULSI) technology, particularly lithography, has led to an aggressive scaling of metal-oxide-semiconductor transistors (MOSTs) to smaller dimensions to increase the packing density (transistors/chip) and to integrate more complex circuit functions onto a single silicon chip or die [I]. The small transistor size also greatly improves the speed of the VLSI/ULSI circuits. For instance, in the 1980's, a typical channel length in the MOST VLSI circuit is of the order of 1gm [2-3]. Currently, MOSTs fabricated with 0.25gm to 0.35tgm channel length are being used to manufacture high density 64Mb DRAM and fast 200 MHz microprocessor with 107-108 MOSTs [4-6]. Recently, MOSTs with 0. 1gm channel length fabricated at research laboratories have been reported [7-10].

Constant-field and constant-voltage scalings are the two limiting approaches used for scaling the MOST to increase transistor density and to improve circuit performance [8, 11-14]. Constant-field scaling, first introduced by Dennard of IBM in 1974 [11], proposed a linear reduction of all the physical geometries of the transistors (oxide thickness, channel length, channel width, and junction depth) and supply voltage with increased doping concentration to scale MOST to smaller features. This approach attempts to maintain a long-channel characteristic in scaled MOST and also to keep the



1






2


drain electric field and power density constant after scaling. However, the present transistor scaling trend has followed the constant-voltage scaling law more closely in order to be compatible with power supply voltage and noise margin constraints. But, scaling down the gate oxide thickness and channel length at constant supply voltage unavoidedly increases the oxide and channel electric fields. These high fields reduce MOST reliability because they cause (1) acceleration of hot carriers (electrons or holes) to high kinetic energies, (2) injection of hot carriers via surmounting the SiO2/Si energy barrier into the gate oxide, (3) quantum-mechanical band-to-band and band-trap-band tunneling of thermal carriers into the gate oxide, (4) charging and generation of oxide traps by the injected carriers, and (5) generation of interface traps [15-28]. The physical charging-discharging and generation-annihilation of the oxide and interface traps by the hot carriers have been identified to have detrimental effects on the stability and reliability of the silicon MOSTs. These hot carriers effects cause a gradual degradation of the transistor electrical characteristics as reflected by changes in transistor's threshold voltage (AVGT), drain-current driving capability (AIDsat), subthreshold distortion (AS) and an increase in standby power consumption. When the degradation of the MOST exceeds an acceptable level in a VLSI/ULSI circuit, for example AVGT=10 mV or AIDsat/IDsat0=-5 %, the circuit may not operate properly and reliable operation of the electronic system using this circuit will be adversely affected. Therefore, the hot carrier effects in deep submicron MOST must be understood and minimized such that the scaled-down MOSTs can continue to achieve a reliable operation Time-to-Failure (TTF) of 10 years. Innovative hot-carrier-resistant transistor designs and careful processing






3



steps have been introduced to minimize the hot-carrier-induced device degradation [2932].

To evaluate the reliability of submicron MOST and its operation TTF without actually operating the MOST continuously for 10 years, the semiconductor industry has adopted a voltage-accelerated stress approach to extrapolate the TTF of a new or production technology [33-37]. In this engineering approach, MOST fabricated from a particular technology is stressed at higher voltages to accelerate its degradation rate. Then the high-voltage accelerated-stress results are empirically extrapolated to operating condition such as 5 V or 3.3 V. This engineering approach is highly empirical and gives unreliable TTF extrapolation to lower voltages as it does not consider the physical dependence of the device failure rate on the kinetic energy of the hot carriers and the threshold energy for a particular denominate degradation mechanism and pathway [38, 23]. In addition, an empirical failure model with fudging or adjustable parameters does not provide any insight to the actual intricate degradation mechanisms and pathways. Thus there is a need for a physics-based TTF extraction algorithm which can provide fundamental physical parameters for evaluating a developing or production technology and also for basic understanding of the basic physics of hot-carrier-induced degradation in MOSTs.

Recently, Neugroschel and Sah [39-40] demonstrated a novel current-acceleration (as opposed to voltage-acceleration) TTF extraction methodology to stress a Bipolar Junction Transistor (BJT) at low emitter-base reverse-bias voltages. The currentacceleration stress technique uses increased hot carrier density at low stress voltages to






4



accelerate the degradation rate of the BJTs. We will demonstrate the extension of this current-accelerated BJT stress methodology to MOST under channel hot carriers (CHC) stress. For MOST, the density of the hot carriers is increased by forward-biasing the substrate/body or source/body n/p junction during CHC stress This approach will be known as current-accelerated channel hot carriers stress (CACHC). Results will be analyzed using a physic-based degradation model.

The methodology adopted in this work will be based on Stress-And-Measure (SAM) experiments "Stress" refers to conventional substrate hot carrier (SHC) and CHC stresses or the new CACHE stress proposed and demonstrated in the following chapter. "Measure" refers to electrical characterization of the device degradation. This SAM methodology will be described in Chapter 2 where an overview of substrate and channel hot carriers stresses is presented. Chapter 3 presents a novel direct-current current-voltage (DCIV) technique, proposed by Neugroschel and Sah [41], to characterize hot-carrier-induced degradation. This technique employs the surface recombination current of one of the forward-biased and gate-controlled p/n junctions of the BJT and MOST structures as the degradation monitor. This measurement configuration is that of a short-circuited-collector BJT and the base current, IB, is measured as a function of the gate-base dc voltage. This DCIV technique was applied to separate the stress-generated oxide and interface traps in submicron MOST which will be described in Chapter 4. Chapter 5 extends the DCIV technique to spatially profile the stress-generated interface trap distribution in MOST. Chapter 6 presents an experimental correlation of the conventional measured electrical parameters (AIDsat,






5


AVGT and AS) with the stress-generated surface recombination current in submicron nMOST.

Chapter 7 presents a new interface trap generation model recently formulated by Sah [42]. This interface generation model will be used to analyze the currentaccelerated channel hot carriers results presented in Chapter 8. Analysis will show that the Si-H bonds at the SiO2/Si interface are broken by hot holes in n-channel MOST during channel hot electron stress and in n/p/n BJT during emitter-base reverse-bias stress, and by hot electrons in p-channel MOST during channel hot hole stress. The broken Si-H bond forms the dangling silicon bond which is the electrically active interface trap.

Chapter 9 delineates four fundamental interband Auger and impact generation pathways in silicon-based MOST where a 4.25 eV hot hole, initiated by primary energetic electron, can be injected over the Si02/Si barrier into the gate oxide to cause positive oxide charging in SiO2. Three distinct electron threshold kinetic energies are found using energy conservation law: (1) 5.37 eV for hot-electron/hot-hole interband impact generation, (2) 4.25 eV for both hot-electron/thermal-hole impact collision pathway and hot-electron/thermal-hole exchange recombination pathway, and (3) 3.13 eV for hot-electron-thermal hole Auger recombination pathway. Experiments are described which verified the existence of the 5.37 eV threshold energy for interband impact generation pathway.

Chapter 10 presents for the first time an interface trap reduction process, observed using DCIV measurement, in p-channel MOST during conventional channel-hot-hole






6


stress. A hydrogen-diffusion model will be proposed to account for the reduction of interface traps in the channel.

Chapter 11 gives the summaries and concludes this thesis.















CHAPTER 2
STRESS-AND-MEASURE (SAM) METHODOLOGY

2.1 Introduction

In a Stress-And-Measure (SAM) experiment, the degradation characteristics of a MOS transistor (MOST) subject to either substrate hot-carrier (SHC) or channel hotcarrier (CHC) stress are monitored by the following electrical measurements: (1) draincurrent vs gate voltage (ID-VG), (2) saturation drain current vs gate voltage (IDsat-VG),

(3) Direct-Current Current-Voltage (DCIV) or IB-VGB, and (4) forward-biased gated emitter-base p/n junction (IB-VBE). These stress and measure steps are automated via IEEE-488 bus in a SAM station. In this chapter, the SAM station setup and the measurement and stress bias configurations are described. An overview of substrateand channel-hot electron stress is also presented.

2.2 SAM Setup

Figure 2.1 shows a block diagram of the SAM station setup. It consists of three parts: (1) a Device-Under-Test (DUT) box, (2) an instrumentation rack of IEEE-488 bus controlled power supplies, voltage-current pulse generators, digital current- and voltmeters, and (3) a Digital Equipment Cooporation MicroVax-II computer. The MOS transistor (DUT), fabricated on a 6" or 8" wafer, is loaded into a light-shielded wafer probe station where the Drain (D) Gate (G), Source (S), Body (X or B) and Substrate (Sub or E for Emitter) of the DUT are connected via the probes and shielded low



7





8










Source (S)

Gate (G) IEEE488 Device- Interface bus DEC
Under- Drain (D) Instrument o MicroVax II
Test Computer
(DUT) Body/Base (B)

Substrate (E
ram of the Stress-and-Measure SAM Station.




Figure 2.1 Block diagram of the Stress-and-Measure SAM Station.






9



leakage-low noise 500 cables to the instrumentation rack. Data acquisition and instrumentation control are automatically controlled by the SAM programs in the MicroVax-II computer through a IEEE-488 interface bus. All the SAM control programs were coded in the VAX-FORTRAN language. A flow chart of the SAM control program is summarized in Figure 2.2.

On the instrumentation rack, a 6-1/2 digits high resolution HP34401 digital voltmeter is used to measure the voltages supplied by three programmable voltage sources: two Keithley K230 and one Hewlet-Packard HP6106 power supplies. The Keithley picometers (K485 & K486) and HP3478 multimeters are used to measure the d.c. currents during stress and measurement cycles (see Figure 2.3). The measurement and stress configurations are multiplexed by a HP3488 scanner and a custom-built Reed switching relays box. The HP3488 scanner contains two relays cards: (1) a 10-channel array of single-pole/single-throw relays controls the Reed relay box, and (2) a 10channel array of double-pole/single-throw relays is multiplexed together with the HP34401 voltmeter to measure the terminal voltages applied on DUT.

2.3 Electrical Measurement

2.3.1 Drain-Current Measurement

Two drain current measurement configurations are implemented by the SAM station: (1) linear ID-VGs, and (2) saturation IDsat-VDS. The difference between the linear and saturation drain current measurement depends on how the drain and gate are connected electrically. Figure 2.3 and Figure 2.4 illustrate the schematic of the SAM station setup for linear and saturation ID measurement, respectively. During the linear






10




Start SAM expt Input Measurement and Stress conditions




Measurements? FhNo Input previous tt,,s and fluence data Stress


ISet stress duration


-- Measurement




No
End?

Yes

Stop SAM expt




Figure 2.2 Flow chart of the SAM control program.










K230
HP3488

HP3478 Body Source K4 .
K486
Gate Drain




N+ N+


P-well

N-epitaxial layer

N+ substrate


Substrate




Figure 2.3 SAM schematic setup for linear ID-VG measurement.






12



HP3488

HP3478 Body Source I
K486 K230 Gate Drain








P-well

N-epitaxial layer

N+ substrate


Substrate




Figure 2.4 SAM schematic setup for saturation IDsat measurement.






13



and saturation ID measurement, the source, body (or p-well) and the substrate of the MOST are grounded. For the linear ID-VG measurement, a programmable voltage K230 is supplied by K230 to give a low (<0.5V) and constant voltage across the drain and source, VDS, and another K230 voltage source is connected to the gate to step the gate voltage, in small incremental step, from accumulation to inversion. The drain current is measured by a sensitive pico-ammeter (K486). The voltages applied to the gate and drain terminals are measured by the HP34401 digital voltmeter. Thus, the ID-VG characteristics of the MOST is obtained by measuring the drain current for a range of gate voltages.

During the saturation IDsat-VD measurement, the gate and drain are tied together and connected to a single K230 voltage source (see Figure 2.4). This connection ensures that the MOST is operating in the saturation mode. The drain current IDsat can either be measured by the K486 picoammeter or the HP3478 ammeter. This flexibility is necessitated by the 2mA current compliance limit of the K486 picoammeter. The voltage supplied by the K230 is measured by the HP34401 voltmeter. By measuring the drain current for a range of drain voltages, the saturation IDsat-VD characteristics can be obtained.

2.3.2 DCIV Measurement

The DCIV measurement, acronyned by Sah [41], is a plot of the d.c. body (or base) current versus gate voltage with a forward-biased p/n junction supplying the minority carriers to recombine with the majority carriers at recombination centers at the SiO2/Si interface of the MOST. In this section, we will only describe the basic measurement






14



K230
HP3488

K485 HP3478

Body Source K486 K230
Base Gate Drain





N+ N+


P-well

N-epitaxial layer

N+ substrate


Substrate




Figure 2.5 SAM schematic setup for DCIV measurement showing the Drain-emitter
DCIV configuration.






15



procedure. Details of DCIV measurement configurations and its basic principles will be elaborated in Chapter 3. Figure 2.5 shows the schematic setup for Drain-Emitter DCIV measurement. A K230 voltage source (constant VEB) is connected to the drain to forward bias the drain n+/p junction. Another K230 voltage source (VGB) is connected to the gate to modulate the Si surface from accumulation to inversion. A K485 picoammeter is connected to the p-well to measure the body current at each VGB step. The source is grounded during the entire measurement.

2.3.3 Gated IB-VB Measurement

The IB-VBE measurement is identical to the standard method used for bipolar transistors' characterization. The only difference here is that a constant d.c. voltage is applied to the gate of the MOST during the IB-VBE measurement. In the case of a standard 3-terminal BJT structure, there is no gate terminal available. The IB-VBE measurement uses exactly the same DCIV setup shown in Figure 2.5. During the measurement, the emitter-base (for example the n+drain/p-body-well junction is forward-biased and the VBD is increased with small incremental steps while a constant voltage is maintained at the gate. The base (or body ) current is monitored using the K485 picoammeter.

2.4 Substrate and Channel Hot Electron Stress

In silicon, thermal electrons at the conduction band edge can gain kinetic energy by tranversing through a potential drop and becomes energetic and hot. These hot electrons are characterized by an effective temperature, Teff = (2/3).(Kinetic Energy)/kB, which is much greater than the lattice temperature, TLattice. For example, if






16



KE= 1 eV, Teff = (2/3).300K(leV/0.025eV) = 8,000 K >> Tlattice= 300 K at room temperature. In nMOST, there are two important hot-electron effects: (1) the substrate hot-electron (SHE) effect where the primary source of hot electrons is due to the current from the substrate, and (2) the channel hot-electron (CHE) effect where the primary source of hot electrons is due to the surface channel current. To investigate these two hot-electron effects in silicon MOST, the stress mode implemented on the SAM station is carefully designed by considering the desired stress conditions and configurations. These will be described in the following two sections.

2.4.1 Substrate Hot Electron Stress

In this work, a substrate hot electron injection (SHEi) technique is used to uniformly inject electrons from the silicon substrate into the gate oxide of the MOST. This injection technique employs a forward-biased vertical bipolar transistor [16] and has been previously used to employed to study the basic degradation mechanisms of the SiO2 gate oxide and its fundamental electrical properties [20, 22, 24, 43]. Optical hot electron injection (OHEi) [44] is a variation of SHEi. The major difference between OHEi and BiMOS-SHEi is the source of minority carriers (electrons) in the p-type silicon body of the nMOST. For OHEi, the hot electrons comes from interband optical generation of electron-hole pair in the surface space-charge-region in the p-silicon substrate underneath the gate oxide. For BiMOS-SHEi, the source of hot electrons is the minority carriers (electrons) injected by the forward-biased n+/p emitter-base junction. BiMOS-SHEi is employed in this work because of its easy and straightforward implementation. For OHEi, a special test structure with thin layer of






17



conductive and transparent gate material or a special window on the substrate surface is necessary. One key advantage of using SHEi is the independent control of three basic stress parameters, the oxide electric field, the carrier kinetic energy and the hot carrier current. The areally uniform electron injection into the gate oxide during SHEi also simplifies the analysis.

Figure 2.6 shows the schematic SHEi setup for a nMOST with junction-well isolation. During the injection stress, the gate of the nMOST is connected to a K486 picoammeter in series with a K230 voltage source. The K486 picoammeter measures the gate current for fluence calculation. The K230 voltage source supplies a positive voltage to the gate for maintaining a constant electric field across the gate oxide. The source and drain are tied together and connected to another K230 voltage source where a reverse bias is applied between the source/drain (collector) and the p-well (base) to bias the vertical parasitic bipolar transistor into active region as well as to control the amount of the band bending in the silicon surface space-charge layer. A forward-bias is applied between the p-well (base) and the substrate (emitter) to supply the minority carriers (electrons). A lkQ resistor was connected in series with the emitter to limit the emitter current. The emitter and collector currents were measured by two programmable HP3478A ammeters. While keeping the reverse-bias between the p-well (base) and source/drain (collector) constant, the forward bias across the p-well (base) and substrate (emitter) was adjusted by a software feedback loop in the SAM control program to maintain a constant gate injection current. The magnitude of the gate





18



K230


SHP3488

K486 HP3478 Body Source
K486 K230








P-well

N-epitaxial layer

N+ substrate


Substrate

1 k9


HP3478


T HP6106




Figure 2.6 SAM schematic setup for BiMOS-SHE injection stress.






19


injection current can be controlled by varying the forward-bias across the emitter-base junction and the reverse-bias across the collector-base junction.

2.4.2 Channel Hot Electron Stress

The channel hot electron stress conditions in nMOST are classified based on the magnitude of the applied VDS and VGS. Two commonly used CHE stress conditions are

(1) VDS < VGS or the "triode" range, and (2) VDS 2 VGS or drain current-saturation range. When VDS < VGS, the transistor is operating in the linear range. The channel electrons are accelerated by the large channel electric field in parallel with the SiO2/Si interface near the drain to gain more kinetic energy. Near the drain junction, the electron trajectory may be deflected sideways towards the SiO2/Si interface by phonon scattering. If the kinetic energy of these impinging electrons is greater than the electron barrier height at the SiO2/Si interface (-3.12eV), they can be injected into the gate oxide. This electron gate current was modeled using the Shockley's "lucky electron model" [45] and expressed [46-48] as

IG 0C Ichexp (-BN/qEMN) (2.1) where Ich is the channel current, EM is the maximum electric field perpendicular to the SiO2/Si interface, and XN is the empirical electron scattering mean-free-path (=78A). The electron barrier height,
OBN = 3.12 2.59x10-4EOX 1/2 10-5Eox2/3 (2.2) where Eox is the oxide electric field near the drain junction. The last two terms takes into account of the Schottky-barrier lowering and the possibility of tunneling through






20


the barrier. This linear CHE stress mode was first employed by Abbas and Dockerty in 1974 to investigate the effect of channel hot electron on the electrical characteristics of the n-channel MOST [49]. Their investigation showed that electrons can be injected into the gate oxide and subsequently captured by the oxide traps by operating the nMOST at the drain current saturation point VD=VG. This bias configuration localized electron trapping near the drain junction and was attributed as the cause of the positive gate shift of the drain current-gate voltage characteristics with increasing injection time. Subsequent experimental investigations and numerical simulations by other researchers have widely supported this localized electron-trapping model under saturation CHE stress [50-52, 55]. Typically, MOST fabricated with poor quality gate oxide (high concentration of electron traps) and with appreciable gate current (high electron injection efficiency) will degrade faster under saturation-point CHE stress [53].

When VDS>VGs, the transistor is biased deep into the drain-current saturation range. Unlike the saturation-point CHE stress, the channel electrons are mainly accelerated by the large electric field in the drain space-charge layer due to VD VDsat instead of by the small channel electric field. If the kinetic energy of the hot electrons is greater than -EG-Si (i.e. VD VDsat > IEGsi/ql, interband impact generation of electronhole pairs occurs. Most of the holes flow into the p-well, giving rise to a body (or well) current in the p-well terminal (Ix). The observed Ix typically peaks at VDs=2VGS [3335, 54-55]. A peak well current will correspond to maximizing the combination of interband impact generation rate of electron-hole pairs and the efficient collection of holes by the p-well. This well. current can overload the substrate-bias generator during






21



operation and cause a snap-back breakdown and CMOS latchup. Due to the 2dimensional electric field distribution in the drain space-charge layer, empirical models with adjustable parameters have been used to model Ix by relating the peak electric field

(EM) at the drain to the drain voltage (VD) [33, 35]

Ix = Ai Ich exp (--B/EM) (2.3) where A and B are empirical fitting parameters. The peak well current was assumed to have an exponential dependent on the maximum electric field at the drain edge. An estimate of the maximum channel electric field was given by [35], VD VDsat
EM = (2.4) L Ydep

where VDsat is the drain saturation voltage, L is the channel length, and Ydep is channel depletion point. Eqn.(2.4) assumes that the maximum longitudinal electric field in the silicon surface layer between the source and drain occurs between the depletion point, Y=Ydep, and the drain metallurgical junction, y=L. The distance from the channel depletion point to the drain metallurgical junction has been empirically modeled by Chan-Po-Hu [56],

L Ydep = 0. 2 ox1/2*XjI/3 (2.5) where xox is the gate oxide thickness and Xj is the drain junction depth.

At large VDS and VGS=VGT, energetic impact-generated electrons and holes can be injected sideway into the gate oxide [57-58]. Between the depletion point and the source, the surface and oxide electric field attracts the channel electrons to the SiO2/Si interface. The surface and oxide electric field is directed in the opposite direction






22



between the depletion point and the drain, attracting the impact-generated holes to the SiO2/Si interface and repelling the electrons away from the SiO2/Si interface. A fraction of the hot holes that gain sufficient energy to surmount the SiO2/Si hole barrier may be injected into the gate oxide, but poor hole injection efficiency is expected due to the larger 4.25eV hole barrier height. Chapter 9 will elaborate more on this channel hot hole injection process.

Two types of CHE stress configuration are employed here: conventional CHE and current-accelerated or CACHE. Figure 2.7 illustrates the schematic setup for conventional CHE stress. During conventional CHE stress, the source and body are grounded. Two K230 voltage sources are separately connected to the gate and drain. The drain current is usually greater than 2mA for wide and short MOST. Thus the HP3478A multimeter has to be used, instead of the K486 picoammter, to measure the drain current. A K486 connected in series with the gate measures the gate current during stress.

For current-accelerated or CACHE stress, the source/body or substrate/body junction is forward-biased to supply additional hot electrons to increase the hotelectron-induced degradation rate of the MOST. Thus, the drain and gate setups are identical to conventional CHE stress. Figure 2.8 depicts the schematic setup for current-accelerated CHE stress using the forward-biased source/body junction. Two K230 power supplies are connected to the drain and gate terminals and referenced to the grounded source. A positive voltage, supplied by a HP6106 power supply, is connected to the body of the n-channel MOST to forward-bias the p-body/n+source junction.






23



Figure 2.9 shows the second current-accelerated which forward-biases the n+substrate/p-body junction. In this case, a negative voltage is applied to the n+ substrate of the n-channel MOST and the p-body and n+source are grounded during CACHE stress.





24



K230


HP3488 K486 I HP3478 Body Source
K486 ~K23( Gate Drain





N+ N+


P-well

N-epitaxial layer

N+ substrate


Substrate


HP3478 T HP6106




Figure 2.8 SAM schematic setup for current-accelerated CHE stress using the
forward-biased source/body junction.





25



K230


HP3488

K486 I HP3478 Body Source
IK486 K230 Gate Drain







P-well

N-epitaxial layer

N+ substrate


Substrate


HP3478 HP6106




Figure 2.9 SAM schematic setup for current-accelerated CHE stress using the
forward-biased substrate/body junction.















CHAPTER 3
DIRECT-CURRENT CURRENT-VOLTAGE (DCIV) CHARACTERIZATION TECHNIQUE

3.1 Introduction

In the MOST with a n/p junction isolation well (usually the pMOST of the present production CMOS chips), the vertical or lateral BJT action is attained by forwardbiasing either the n+/p well junction or the source/well or drain/well p/n junction. In the MOST without the n/p junction isolation well (usually the nMOST of the present production CMOS chips), only the lateral BJT action can be attained by forward-biasing the drain/body or source/body p/n junction.

In this work, a simple and high sensitivity Direct-Current Current-Voltage (DCIV) technique is employed to characterize the hot-carrier-induced degradation in silicon metal-oxide-semiconductor transistors (MOSTs). In this chapter, the basic measurement technique is discussed. This DCIV technique is based on the gate voltagecontrolled electron-hole recombination rate at the SiO2/Si interface traps. This technique was first employed by Sah in 1961 to study the surface recombination and channel on silicon transistors [59-60]. Recently, this technique has been re-introduced by Sah and Neugroschel to study the effect of the stress-created oxide (NOT) and interface (NIT) traps on the reliability of advanced MOS and Bipolar junction transistors (MOSTs and BJTs) [39-41, 61-63, 96].




26






27



3.2 Measurement Configurations

The DCIV technique is a gate-controlled BJT measurement and employs the lateral or vertical BJT of the silicon MOST structure. This was once referred to as BiMOST [41] which will be discarded and replaced by the above general description [63, 71]. Figure 3.1 shows a cross-sectional view of a nMOST illustrating the vertical and lateral BJTs. Figure 3.1 illustrates three n/p junctions that can be forward-biased the emitter to inject minority carriers into the p-base well: (1) the Drain/Base (Body) n+/p junction,

(2) the Source/Base (Body) n+/p junction, and (3) the Substrate/Base (body) n+/p junction. In this work, the body and base are used interchangeably in the discussion with the same subscript, B or b. Thus, there are three basic DCIV measurement configurations: two top-emitter (1) Drain-Emitter=DE, or (2) Source-Emitter-SE, and

(3) bottom-emitter=BE (Substrate-Emitter is not used to avoid confusion with SourceEmitter abbreviation). In the DE-DCIV configuration, the n+drain/p-body is the forward-biased emitter junction and the n+source/p-body is the short-circuited n+collector/p-base junction. Similarly, in the SE-DCIV configuration, the n+source/p-body junction is the forward-biased emitter junction and the n+drain/p-body is the short-circuited n+collector/p-base junction. In the BE-DCIV configuration, the n+substrate/p-body is the forward-biased emitter junction and the n+drain/p-body and/or n+source/p-body is/are the short-circuited n+collector/p-base junction(s). This systematic classification, acronyned by Sah [63, 71], of the DCIV measurement configurations is necessary to exemplify the physical location of the CHC-induced degradation, as will be discussed later.





28








Source Gate Drain Base

SiO2 QT Qt
PIT .OC lol
n+ n+ p+base contact

p-base well n-epitaxial layer n+ substrate


Substrate




Figure 3.1 Cross-sectional view of a n-channel MOST showing the lateral and vertical
parasitic n/p/n BJT.






29


During DCIV measurement, the emitter-base junction is forward-biased and the collector-base junction is short-circuited. The gate-base voltage is stepped from accumulation to inversion. The base or body current Ig is measured using a picoammeter at each gate voltage step. Thus, a DCIV measurement is a plot of the base or body terminal current versus the gate-base or gate-body voltage at a constant forward-biased emitter-base voltage.

3.3 Components of IB

During hot-carrier stress, the interface traps (open triangles in Figure 3.2) along the SiO2/Si interface can be generated over the drain/body junction space-charge-region and over the n-channel region from hot carriers breaking the strained or weak hydrogen bonds (Si:H and SiO:H). Thus, during DE-DCIV measurement, the measured IB-DE contains two emitters (not shown in Figure 3.2) and six base recombination components [59-60, 63-65]. The two emitter components are the electron-hole (e-h) recombinations at the recombination centers in the drain's bulk-quasi-neutral region (IB-el-bulk) and at the interface between poly-silicon and crystalline-silicon contacts (IB-el-contact), respectively. The base components are due to e-h recombinations centers or interface traps located in the six regions. In Figure 3.2, these six base recombination components are distinguished by one alphabetic and one numeric subscript: c-channel region, s-surface space-charge region, b-bulk space-charge region; I and 2 from IB o exp(qVBE/nkBT) where n=l for Shockley's exact p/n junction diode law and n=2 for the simplified Sah-Noyce-Shockley (SNS) space-charge-layer recombination law [59, 65]. These six components are: (1) IB-DE-c-l over the n-channel-interface, (2) 'B-DE-s-2 over






30












+ VBE=
VGB 0.4V
Body Source Gate Drain Base Collector Gate Emitter




n+ n+


p-Body O(3) IB-b2 (4)I (5)IB-2.
p-Well

222CTS p+Substrate




Figure 3.2 Cross-sectional view of n-channel MOST biased in the Drain-Emitter (DE)
DCIV configuration. Six base current recombination pathways are depicted. White squares and hexagons in SiO2 are oxide electron and hole traps. White and black triangles are active (Si. and SiO*) and inactivehydrogenated (Si:H and SiO:H) interface traps. Square-grid-shaded areas are space-charge regions of p/n and p+/p junctions in Si. Black and white circles are electron and holes in Si. White squares in Si are electron-hole
recombination-generation centers. Adapted from Sah [63].






31


the surface space-charge region of the drain/body junction along the n-channelperimeter, (3) IB-DE-b2 in the drain/body junction bulk space-charge region, (4) IB-DE-bl in the drain/body junction bulk-quasi-neutral region, (5) IB-DE-b2', same as (1) but outside of the n-channel-perimeter, and (6) IB-DE-cl', same as (2) but outside of the nchannel-perimeter. Thus, the measured IB-DE is a sum of the 8 components described above.

IB-DE = 'B-el-bulk + IB-el-contact +

IB-DE-cl + IB-DE-s2 + +IB-DE-b2 + IB-DE-bl +

IB-DE-b2' + IB-DE-cl' (3.1)

Of all the base-current components described in (3.1), only IB-DE-cl and IB-DE-s2 can be modulated by the gate voltage [59] and are sensitive to oxide charge QOT built-up and interface traps NIT generation during hot-carrier stress. Thus the 6 gate-voltage independent components can be subtracted. The above systematic classification of the IB components during DE-DCIV measurement also applies for SE-DCIV.

Figure 3.3 shows the components of IB in the BE-DCIV or Bottom-Emitter/DrainSource-Collector DCIV measurement configuration for a nMOST fabricated with a n+/p junction well (in most present production chips, only the pMOST has the p+/n isolation well junction). During BE-DCIV measurement, the collector-base junction is either grounded or weakly reverse-biased. Thus, the electron-hole recombination components along the surface space-charge-region and in the bulk space-charge region of the collector/base (drain/body and source/body) junctions are negligible due to low electron and hole concentrations in these regions. In fact, thermal generation of electron-hole






32













SVGB
Body Source Drain
Base Collector Collector
B 1C n+poly Si

Sio2 SiO2 Nor N,. SiO2 NT Nrr:H SiO2


6) Isco
p-Body (4)IB-b
p-Well

9 70 n+Substrate

Substrate VBE= -0.4V I Substrate





Figure 3.3 Cross-sectional view of n-channel MOST biased in the Bottom-Emitter
(BE) DCIV configuration. Three base current recombination pathways are
depicted. See Figure 3.2 caption for the description of symbols.






33



pairs prevail in these regions but can be neglected because the large base recombination current will overwhelm this small generation current even at small forward bias [96]. Thus, the measured Ig in the BE configuration consists of only two emitter components and three base recombination components. The two emitter components are the e-h recombinations at the recombination centers in the n+-substrate's bulk-quasi-neutral region (IB-BE-el-bulk) and at the interface between the n+ substrate and metal contacts (IB-BE-el-contact), respectively. The three base e-h recombination components are: (1) IB-BE-c-i over the n-channel-interface which is same as IB-DE-c-1, (2) IB-BE-b2 in the n+substrate/body junction bulk space-charge region, (3) IB-BE-c-I' same as (1) but outside of the n-channel-perimeter. Thus, the measured IB-BE is given by

IB-BE = IB-BE-el-bulk + IB-BE-el-contact + IB-BE-cl

+ IB-BE-b2' + IB-BE-cl' (3.2) In (3.2), only IB-BE-cl is modulated by the gate voltage. Thus the BE-DCIV measurement configuration gives greatly simplified data analysis and provides the unique locationing of the NIT (n=l) in the channel region. This BE configuration had been used by other researchers to study the surface recombination rate or velocity So at the SiO2/Si interface [66, 67] and the areal uniformity of positive oxide charge build-up and interface trap generation using substrate hot-electron injection [61].

3.4 Theory of Surface Recombination Rate

The surface recombination current components discussed in section 3.3 can be evaluated by extending the Shockley-Read-Hall (SRH) recombination kinetic for a single trap level [68] to determine the surface recombination rate Rss of electron and






34



hole through the interface trap which has a discrete energy level ET in the silicon energy gap. The following presentation is based on the lectures given by Sah [69]. Consider an infinitesimal interfacial layer of thickness 8x at the SiO2/Si interface where the electron-hole recombination occurs (Figure 3.4), the steady-state electron-hole recombination rate Rss (cm-2s-1) at the interface traps, in the shaded area region, is CnsCpsNP enseps
R, = NIT(x) bx dx (3.3) CnsNs + ens + CpsPS + ep CnsCpsNSPS enseps
NIT (3.4)
cnsNs + ens + cpsPS + ep where cns and cps are the electron and hole capture rate coefficients (cm3sec-l), ens and eps are emission rate coefficients of the trapped electron and hole (sec-1), Ns and Ps are the concentrations of electron and holes at the interface (cm-3), and NIT is the areal density of interface trap (cm-2). At and near thermal equilibrium (i.e. neglect electric field or hot carrier effect) ens can be related to cns by

ens ET EI
ni exp( ) = nj (3.5) Cns kBT

Similarly for eps and cps,

eps EI ET
ni exp( ) = p1 (3.6) Cps kBT

In Eqns.(3.5) and (3.6), ni is the intrinsic carrier concentration of silicon, El is the intrinsic energy level, kB is the Boltzman constant and T is the ambient temperature.






35
















Si02 Si





interface traps in this thin interfacial layer











6x



Figure 3.4 Cross-sectional view of a thin interfacial layer at the SiO2/Si interface.






36



Substituting Eqns.(3.5) and (3.6) into Eq.(3.4) and using nlP1 = ni2, Rss can be simplified as

(NsPs ni2)
Rs NIT (3.7) Cps- (Ns + n1) + ns-l(P + )(3.7)

The surface recombination current IBS is then the areal integration of the surface recombination rate Rss along the y- and z-direction.


IBS q J Rs, dy dz (3.8)


Equations (3.4) and (3.7) show that there are three factors that will affect the characteristics of the surface recombination currents. They are: (1) the spatial interface trap density, NIT(y,z), (2) the spatial surface electron and hole carrier densities, Ns(y,z) and Ps(y,z), and (3) the electron and hole capture and emission coefficients, cns, Cps, ens and eps, and the energy level of the interface trap in the silicon energy gap, ET. During hot carrier stress, an increase in the density of the stress-generated interface traps will increase the magnitude of the surface recombination currents. This is the basis of the DCIV which employs the surface recombination component to monitor hot-carrierinduced degradation in MOSTs. The capture and emissions coefficients are assumed not to be affected by the hot-carrier stress.

Several salient characteristic properties of the surface recombination rate Rss are best understood by considering the IB-cl components, shown in Figures 3.2 and 3.3. In the following, we will use the IB-cl component, measured using the Bottom-Emitter DCIV configuration, as an example in our discussion. Consider the transition energy






37


band diagram of a nMOST shown in Figure 3.5 biased in the bottom-emitter DCIV configuration, the gate voltage modulates the amount of surface energy band bending

(Vs) at the SiO2/Si interface and the forward-biased bottom emitter injects minority carriers (electron) into the p-base well. The rate of electron-hole recombination at the interface traps, Rss, will depend on the surface electron and hole concentrations, Ns and Ps. As indicated in Figure 3.5, they are given by


Ns = NEE exp exp(F (3.9)



ni exp I (V v (3.10) kBT

q (Vfs VF + VBE
ni exp k(3.11) kBT

and

-qys
PS = PE exp ( ) (3.12) kBT

In Eqns. (3.9) and (3.11), NE and PE are the equilibrium electron and hole concentrations. Low injection level is also assumed, i.e. N and P < PB=PE= niexp(qyFl/kT).

Figure 3.6 illustrates the relation between the surface recombination rate and surface potential at the SiO2/Si interface by plotting Rss versus Vs for four different forward-bias voltages. The shape of the Rss curves is the same but the magnitude of Rss depends exponentially on the excess surface carrier (electron) density. The surface






38











qAsI F E < ** F -TF qVE B q(VGB-Vl B E Ec E Ev
----------qVo



I I
Space-charge layer

n+ Poly Si SiO2 p-Si n+ Si
Gate Oxide I Base Emitter





Figure 3.5 Cross-sectional view of the transition energy band diagram.





39







1015

VBE=0.5V 14

C/ 0.4

E 0.3V



CE' 1011


1010
-0.1 0.0 0.1 0.2 0.3 0.4 0.5 Vs/(l V)


Figure 3.6 Relation of Rss and Vs with VBE as parameters. Assumed mid-energylevel interface trap in Si energy gap and Cns=cps. NAA= 1016 cm-3. T=300
K.






40



potential at which the recombination rate peaks is negatively displaced as the surface minority carrier (electrons) density increases due to the injection of minority carriers from the forward-biased emitter. The surface potential vs-max corresponds to peak surface recombination rate is [59]

Vfs-max = VF VBE/2 (kBT/2q) .log(cps/cns) (3.12) At this surface potential, the surface is intrinsic, i.e., the electron and hole concentrations are the same and equal to qVBE
Ns = Ps = ni exp( ) (3.13) 2kBT

Assuming ET EI = 0, co = cns = cps and Ps = Ns > ni, coNITni qVBE
Rss = ( ) exp( ) (3.14)
2 2kBT

This simplification implies that under maximum recombination condition, IB-BE-cl o exp(qVBE/2kBT). Next, consider the case for Ps > Ns > ni,

Rss = co NIT Ns (3.15) qWs-VF+VBE
= coNITniexp ( ) (3.16) kBT

Here, IB-BE-cl oc exp(qVBE/lkBT).

3.5 Relationship between '- and VGB Under Non-Equilibrium Condition

So far, we have only considered the relationship of Rss as a function of Vs. It is necessary to be able to relate the applied gate voltage (VGB) during DCIV measurement to the surface band-bending potential sW, This can be obtained by modifying the boundary conditions when solving the Poisson Equation, Eqn. (3.17), for w.






41



Esi (dE/dx) = -si (d2V/dx2)

= q( P N PB + NB ) (3.17) where Esi is the silicon dielectric constant, E is the electric field strength into the bulk Si, P is the hole carrier concentration, N is the electron carrier concentration, PB is the bulk hole concentration, and NB is the bulk electron concentration. The boundary conditions at the SiO2/Si interface and in the silicon bulk are given as:

V(x=O) = V/s (3.18)

V(x==) = 0 (3.19) E (x=0) = Es (3.20) E (x=) = 0 (3.21) P (x=O) = Ps (3.22) = PBexp ( ) (3.23) kBT

P (x=m) = PB (3.24) qVF
= niexp ( ) (3.25) kBT

N(x=0) = Ns (3.26) qVfs
= NBexp ( ) (3.27) kBT

N(x=~) = NB (3.28)
-qVF qVBE
= niexp ( )exp ( ) (3.29) kBT kBT

Using the above boundary conditions and solving Eq.(3.17) by quadrature,

d2V/dX2 = (dV/dx) (d/dV) (dV/dx) = (1/2) (d/dV) (dV/dx)2

= (1/2) (dE2/dV)2 (3.30)






42



The surface electric field Es can be obtained as

E= E(x=0) = E (Vs) (3.31) = (2kBTPB/ESi) *


[exp (-Us)-1+Us] + [exp (Us) -1-Us) ] exp (-2UF+UBE) where Us, UF and UBE are the normalized potential with respect to kBT/q (Us = qs/kaT, UF = qVI/kaT, and UBE = qVBE/kBT).

The relationship between the applied gate voltage VGB and the surface bandbending potential ys is obtained by using the MOSC voltage equation.

VGB VFB V'S = Vox = siEs/Cox (3.32) where VFB is the flatband voltage, Vox is the oxide voltage, and Cox is the oxide capacitance per unit area. Thus, for a given VGB, Eq.(3.32) can be solved iteratively for vs. Three exact asymptotic solutions for Eq.(3.32) can be derived by weighting the relative magnitude of the carrier concentrations at the SiO2/Si interface under surface accumulation, depletion, and inversion conditions [70]. Under surface accumulation condition where Us <0, (VGB-VFB-Vs) 2
s = (kBT/q) loge + AA (3.33) (4kBT/q) V,

Under surface depletion condition where 0< Us < 2UF UBE, (VGB-VFB- (kBT/q) AD
V's = VGB-VFB+2VAA 1- 1 + (3.34) Under surface inversion condition where Us > 2U-UVAA

Under surface inversion condition where Us > 2UF UBE,






43



kBT (VGB-VFB1-S) 2
is = 2VF-VBE+ loge[ (VGB-VFB-/)I (3.35) q (4kBT/q) VA

In Eqns. (3.33)-(3.35), the iterative correction terms (AA, AD and AI) are

AA = -Us- [exp (Us) -1-Us] exp (-2UF+UBE) (3.36) AD = 1-exp (-Us) [exp (us) -1-Us] .exp (-2UF+UBE) (3.37) AI = l-(l+Us)exp(-Us) + [exp(-2Us)-1+Us] (3.38)

Using Eqns.(3.31)-(3.33), the relationship between ys and VGB is computed for xox=15nm, NAA=1016cm-3, VFB=OV, and T=300K. The result is shown in Figure 3.8 with VBE as parameter. In the accumulation and depletion ranges, the results computed for non-zero VBE's are similar to the case with VBE=O V. This indicates that, under low-to-moderate forward-biases, the excess surface electron concentration due to electron injection from the forward-biased emitter junction is not appreciable to affect the field and potential distribution in the surface space-charge-layer. However, in the strong inversion range, excess surface electron concentration reduces the amount of surface band bending by VBE.

S, inv = 2VF VBE (3.39)

The results shown in Figures 3.6 and 3.7 are combined and replotted in Figure 3.8 which illustrates the gate-voltage dependence of surface recombination rate, Rss, for four forward-bias conditions (VBE=0.2, 0.3, 0.4 and 0.5V). This theoretical calculations indicates that base current, measured during DCIV, will exhibit a maximum when its surface recombination components dominate. For a single energy-level interface trap,





44







1.0
VBE=O.OV0.8
0.2V

0.6- 0.3V 0.4V



>" 0.2

0.0

-0.2
-2 -1 0 1 2

VGB/(1 V)


Figure 3.7 Relation between surface band-bending potential (Ws) and applied gate
voltage (VGB) under equilibrium (VBE= 0 V) and non-equilibrium (VBE= 0.2, 0.3, 0.4 and 0.5 V) conditions. Parameters used: xox= 15 nm, NAA=
1016 cm-3, VF= 0 V, and T= 300 K.





45







1016


1014 VBE=0.5V
S0.4

1012- 0.3
00.2



IE 108

106 |
-0.5 0.0 0.5 1.0

VGB/(1 V)


Figure 3.8 Relation between surface recombination rate (Rss) and applied gate voltage
(VGB) under equilibrium (VBE= 0 V) and non-equilibrium (VBE= 0.2, 0.3, 0.4 and 0.5 V) conditions. Parameters used: xox=15 nm, NAA=1016Cm-3,
VFB=0 V, N,= 10'0 cm-2 and T=300K.






46


the IB peak occurs when the surface electron and hole densities are comparable to each other.

3.5 Summary

A d.c. current-voltage (DCIV) technique is introduced in this chapter to investigate hot-carrier-induced degradation in MOS transistors. This technique uses the base current IB in a gate-controlled parasitic BJT of a MOST to monitor the stress-generated oxide and interface traps. It has several unique features: (1) it is a pure d.c. measurement which greatly simplifies instrumentation setup; (2) it is directly applicable to submicron area devices in conventional production CMOS-based transistor structure and can monitor the transistor's degradation kinetics under hot-carrier stress; (3) the magnitude of IB is directly proportional to the interface trap density NIT or QIT/q and thus not affected by inhomgenity of NIT; (4) the surface recombination rate Rss and minority injection level can be independently controlled by the applied gate voltage and forward bias voltages, respectively; and (5) it has high sensitivity which gives a minimum measurable NIT of < 109 traps/cm2 (Figure 3.8).

The DCIV technique has been recently packaged into a turn-key diskette (SAMDCIV [71]) for the HP4156 (D.C. Parameter Analyzer), a typical characterization equipment used among semiconductor companies. This diskette, together with a userguide and tutorial documentation, has been distributed to members of the Semiconductor Research Corporation as part of a technology transfer effort.





47






3 l I II Top-Emitter BJT


Z2
c AB= 1pA
0

1
S* AIB =(qASon/2)exp(qVBE/2kT)=l pA
So =2cm/sec = T NIT=109cm-2
0 I I I, , I I I I
-1.0 -0.5 0 0.5 1.0

VGB/ (1V)


Figure 3.9 Sensitivity test of the top-emitter IB measurement of interface traps of an
unstressed BiMOS with nMOST and npn BJT at VEB= -0.3 V and 297 K.
xox= 15 nm. W/L= 100 tm/1.6 um.















CHAPTER 4
SEPARATION OF INTERFACE AND OXIDE TRAPS USING THE DCIV METHOD

4.1 Introduction

Electrically active oxide (QoT) and interface traps (QIT) are known to affect the electrical characteristics of metal-oxide-semiconductor capacitor (MOSC) and metaloxide-semiconductor transistors (MOST). Areally uniform and charged oxide traps in MOSC and MOST will cause the Cg-VG and IDS-VDS characteristics to shift in parallel along the gate-voltage axis since the density of the oxide trap has no functional dependence on the gate voltage, or the surface band bending condition. However, the situation is more complex for interface traps because the density of charged interface traps depends on the electron and hole concentrations at the SiO2/Si interface which are controlled by the applied gate voltage. Hence, interface traps will distort the Cg-VG and ID-VG characteristics.

The traditional measurement techniques high-frequency Cg-VG (input capacitance) and d.c. ID-VG (transfer characteristics) have been routinely used to study the generation and charging of oxide and interface traps during hot carrier stress. For the two-terminal high frequency Cg-VG method, the stress-created oxide trap density is monitored by the change in flat-band voltage while the interface trap density is calculated from experimental data using the Terman method. This method requires a large area MOSC




48






49


test structure for accurate measurement of the small-signal capacitance and thus is not suitable for test structure with very small transistors. For the d.c. ID-VG method, the extracted threshold voltage shift can be used to monitor the density of the oxide traps, provided that the oxide trap density is areally uniform and the interface trap density is sufficiently low. The interface trap density can be qualitatively obtained from the subthreshold slope change. However, if both oxide and interface traps are generated during the hot carrier stress experiment, then it is not easy to separate the relative contribution of QOT and QIT from a single ID-VG measurement because the total gate voltage shift is due to the build-up of both QOT and QIT.

AVG = AVG-OT + AVG-IT (4.1) (AQTT + AQIT)
(4.2)
C,

where AVG-OT is a component of the total gate voltage shift, AVG, caused by the charged oxide trap. By the same argument, the interface trap contributes a AVG-IT component. In addition, ID also depends on mobility which is changed by QIT and QOTIn this chapter, we will show that the DCIV method can be used to separate the AQoT and AQIT [41]. This novel method contains two features: (1) the increase in the magnitude of base current is used to monitor the stress-generated QIT, (2) the lateral voltage shift of the collector current versus gate voltage is used to detect the QOT4.2 Experiment

To demonstrate the separation of QIT and QOT using the DCIV method, a nMOST, fabricated with a junction-isolated p-well, is subjected to Substrate Hot Electron






50



injection (SHEi) to uniformly generate QIT and QOT. The transistor has a drawn channel length of 1.6jtm and a width of 100ptm. The gate oxide, thermally grown at 8500C in a dry oxidation ambient, is 150A thick. A cross-sectional view of the nMOST is shown in Figure 4.1. The stress-generated Qrr and QOT are monitored by the top- and bottom-emitter DCIV, and the conventional ID-VG measurements. The top- and bottom-emitter measurement configurations are shown in Figure 4.2.

Figure 4.3 shows the energy band bending of the gate oxide and the Si substrate in the nMOST during SHE injection. A positive voltage (VG=7.5V) is applied between the gate and the source and drain junctions which are connected together to invert the channel underneath the gate oxide. This inverted n-channel, together with the n+ source and drain junctions, serves as the collector for the vertical n+/p/n-channel bipolar transistor. The p-well is the base and the n-/n+ epitaxial substrate serves as the electron emitter. The emitter/base junction is forward-biased to inject electrons into the p-well base. These injected electrons diffuse upwards through the p-well and are accelerated towards the Si/SiO2 interface by the electric field in the surface space-charge layer of the reverse-biased collector/base n-channel/p-well junction. The reverse bias, VBn, provides the accelerating field for these electrons. When the channel is strongly inverted, VcB will be the potential difference between the hole quasi-Fermi potential Vp=Fp/q, in the p-well and the electron quasi-Fermi potential, VN=FN/(-q), at the Si/Si02 interface. The accelerating electrons may acquire sufficient energy in the Si space-charge layer to surmount the Si/SiO2 barrier (3.12eV). Hence, the minimum VCB that must be applied to accelerate the injected electrons into the gate oxide is about





51










Source Gate Drain Base


7S'0 A IT '3QOT n+ n+ p+base contact
p-base well n-epitaxial layer n+ substrate


Substrate


Figure 4.1 Cross-sectional view of the BiMOS BJT-MOST transistor structure used.






52












Top-Emitter Configuration Bottom-Emitter Configuration
E ov G B C Iov G B
-0.3V F03VOV OV OV
Si0, S OT




__ I LI IT I I p~ l il


n-epitaal layer
n+Collcr I rn itter
ovLc -0.3V E







Figure 4.2 Cross-sectional views of the top-emitter and bottom-emitter DCIV
measurement configurations. The bias used in the experiment are also
depicted in the figure.






53











Electrons

---- ------- Ec
F, - -_-_ - FN [O qVcs

Bulk electron
trap ,

FN qVox Ec F,

Ey I--Space charge region I




n+ Si gate SiO2 p-Si n+ Si






Figure 4.3 Transition energy band diagram of BiMOS substrate hot electron injection.






54



3.12V. In the SHEi experiment, a reverse bias of 4.OV is chosen for VCB. Some of the electrons that are injected into the SiO2 can be captured by the neutral oxygen vacancy centers [72], giving

V0 + e- 4 Vo- (4.3) and results in a net built-up of negative QOT. Because of the high kinetic energy (-4eV from VSB=VDB=4.0V) of the accelerating electrons in the surface space-charge-layer, they can also break the weak or strained interfacial bonds (Si-Si, Si-O, Si-H, SiO-H) via direct impact to create new interface traps [15] since the bond energy of of these interfacial bond is of the order of 3eV.

During SHEi, the gate oxide current is kept constant, within 5% of the initial preset gate current, by digital feedback control of the emitter/base forward bias. The oxide electric field, Eox, is given by (VG Vs)/xox. The gate electron fluence, FG or NINJ, is calculated as follows,

NINJ = S (JG/q) dt (4.4) = (JG/q) tstress (4.5) where JG is the constant gate current density, q is the electronic charge (1.6x 1019 C), and tstress is the total stress time in seconds.

One key feature of using the SHEi technique is that the three basic stress parameters, the electron kinetic energies in the SiO2 and at the SiO2/Si interface, and the injection current can be independently controlled [20, 22, 43]. The electron kinetic energy in the SiO2 is controlled by the voltage applied between the gate terminal and the n+source/n+drain terminals provided the gate voltage is sufficient to invert the silicon






55



surface layer underneath the gate oxide. The electron kinetic energy in the silicon surface space-charge layer is determined by the reverse bias applied between the n+source/n+drain and the p-well. The magnitude of the gate current can be adjusted by the emitter-base forward bias voltage which determines the emitter current, and by the collector-base reverse bias which controls the kinetic energy distribution of the electrons at the Si/SiO2 interface.

4.3 Results and Discussions

Figures 4.4 shows the IB-VGB and IC-VCB curves, measured in both the top-Emitter and bottom-emitter configurations, and the ID-VDS curves measured before and after the SHE stress. In Figure 4.4(a), the increase in the magnitude of the post-stressed IB peak is attributed to an increase in the density of the stress-generated interface trap. The VGB-IT component in Eqn.(4. 1) can then be determined using this stress-induced IB, AIB

- IB(post-stress) IB(pre-stress), using the following relations [74].

AIB = (qAniASO/2) exp (qVBE/2kBT) (4.6) ASO = (7r/2 ) OEth ANIT (4.7) AQIT = -CoAVGBIT (4.8)

SqANT (4.9) = qADIT AEIT (4.10) ASO is the increase in surface recombination velocity due to the stress-generated interface traps. AG is the gate area of the nMOST. ni is the intrinsic carrier concentration. Oth is carrier thermal velocity. In eqns. (4.5) and (4.6), it is assumed that the density-of-state (DIT) and the carrier capture cross sections (an = 0p = ao) of the






56






800

1*- top emitter 600 NN=5xl018cm-2_ stressed a 400 B

Gobottom emitter 200- Nj=1017cm-2 10X
unstressed
(a) 0
140 I I

S.-bottom emitter 120 N- IAV0 C1o


-/ // top emitter
o 100 i unstressed /b It stressed
80i Nm,=5xl018cm-2 80 I FB/

(b) 60

10-3 I



10-6

AVGB= AVGOT + AVGB-IT

10-9
unstressed stressed"/. NN=--5x 018"cm-2

(C) 10-12 . / I .
-1.0 -0.5 0 0.5 1.0

VGB / (1V)



Figure 4.4 Effect of areal-uniform SHEi stress on the n/p/n BJT and nMOST
characteristics as a function of VGB. All stressed at 5x 1018 electron cm-2 fluence except bottom-E IB in (a). Measured at VEB= -0.3 V, VCB= 0 V and 297K. (a) Base current and (b) collector current in the top-emitter and
bottom-emitter configurations. (c) Subthreshold ID-VGB of the nMOST.






57



interface traps in the Si energy gap are independent of energy. In Figure 4.4(a), the IB-peak for the top-Emitter curve stressed at an electron fluence of 5x 10'8cm-2 occurs at a surface band-bending of Vs = VF 0.24V and gives ASo = 1600cm/s. The IB-peak for the bottom-Emitter curve stressed at an electron fluence of 1x 1017cm-2 occurring at s = VF 0.26V gives ASo = 40cm/sec. In the above example, 's for the top-E and bottom-E curves are determined using the relations presented in chapter 3.

In Figure 4.4(b), the pre-stressed Ic-VGB curve is observed to be nearly flat in the accumulation range and sharply increases when VGB > VFB = -0.55V but subsequently saturates to a flat plateau when VGB 2 VTH = +0.05V. These features can be adequately accounted for by considering the relative geometrical increase in the emitter or collector area contributed from the electron channel. In the accumulation range, Ic is only proportional to the area of the n+drain/p-base well junction and thus does not depend on VGB. When VGB 2 VFB, an electron surface channel begins to form and increases the effective collector area that collects the injected electron from the bottom-emitter, resulting in an increase of IC. When VGB > VTH, the p-Si surface is inverted by the gate-induced electron channel. Thus, Ic will be proportional to the sum of the area of the n+drain/p-base well junction and this inverted electron channel. The three characteristic Si surface potentials or Si energy band bending (FB flatband at %s=OV, INV inversion at equal electron-hole surface concentration Ns = Ps or 's = VF VBE/2, and TH = threshold or strong inversion at Ns = PBase or ys = 2VF VBE) are indicated on the pre-stress IC-VGB curve. As seen from Figure 4.4(a), the SHEi has generated substantial interface traps indicated by the large increase in IB. But, the post-






58



stress IC-VGB curve only exhibits a lateral and almost parallel shift from the pre-stress curve with nearly identical height for both the top-Emitter and bottom-Emitter measurement configurations. This lateral positive gate voltage shift AVGB is attributed to the charging of the neutral oxygen vacancy centers as shown in eqn.(4.3). Thus the VGB-OT components in Eqn.(4.1) can be determined from

ANOT = AQIT/q (4.11)

= (CO/q) AVGB-OT (4.12) = (Co/q). [-0.2 (-0.55)1 (4.13) = 4.3x1011cm-2 (4.14)

The effect of hot-electron-induced degradation on the nMOST's characteristics can also be observed from the conventional ID-VGB measurement shown in Figure 4.4(c). After injecting the gate oxide with an electron fluence of 5x1018cm-2, the poststressed ID current decreases with increasing VGB and there is clearly a distortion in the subthreshold regime. The subthreshold slope change AS can be employed to estimate the density of interface traps using the following well-known equation [73].

DIT = (Co/q) (q/2.303kBT)AS (4.15)

= 1012cm-2eV-1 (4.16) However, the ID degradation is clearly due to the stress-generated AQIT and AQOT whose presence is unambigously confirmed by the IB-VGB and Ic-VGB measurement. This is depicted in Figure 4.4(c) where the amount of total gate voltage shift AVGB extracted at a constant drain current is shown to consists of two components AVGB-OT and AVGB.IT. The separation of the total gate voltage shift AVGB into these two






59



components cannot be accomplished from one single ID-VGB measurement alone unless additional properties of the interface traps are known or assumed. However, this can be easily separated using the DCIV technique. On the IB-VGB plot shown in Figure 4.4(a), the post-stressed IB peak increases by about 700 times which translate to about 1012 cm-2 NIT after 5x 1018 electron/cm-2 SHE stress. On the IC-VGB shown in the Figure 4.4(b), the post-stressed IC curve shifted by about 0.35V which translates to a negatively charged QOT density of 4.3x1011 cm2.

Additional examples are given in Figure 4.5(a)-(c) for the bottom-Emitter configuration which use Ic to monitor negative, positive, and turn-around AQoT induced by stress. Figure 4.5(a) is identical to Figure 4.4(b) showing positive AVGB caused by the negative oxide trap -QoT. Figure 4.5(b) shows negative AVGB (curve 2) from positive oxide trap +QoT. The generation of positive oxide trap +QoT is anticipated from the electron-impact emission of electrons trapped by the neutral oxygen vacancy [72] for a SHEi condition with VGB= 12V and VDB=VSB= 10V.

Vo0 + e* -* V0o + 2e- (4.17) Figure 4.5(c) illustrates the QOT turn-around effect as coined by Young [75]. A short CHEi stress (-lsec) at VGB=VDB= 15V with the source floating injects electron into the oxide along the entire length of the strongly inverted n-channel because VGB >> VTH. The electrons are captured by the existing oxide trap resulting in -QoT (Curve 1). After an additional 500 sec stress, there is a built-up of positive oxide traps from the impactemission pathway described by eqn.(4.17) which compensates the -QoT and shifts the curve to the right (Curve 2).






60



140

120


O 100 -0 o I
80

(a) 60 1 1 .
120

100


04




(b) 40 .
120 I . " I 100 -VGB / (1V)
-0.3 V, VCB= 0 V, and 297K. (a) 10 cm-2 SHEi stress at VGB 7.5 V,
o 60

(C) 40 . I . I . I
-1.5 -1.0 -0.5 0 0.5






Figure 4.5 Effect of stress on the collector current measured at VEB (bottom-emitter)=
-0.3 V, VCB= 0 V, and 297K. (a) 5x10'8 cm-2 SHEi stress at VGB = 7.5 V, VCB = 4.OV and IG = InA. (b) 1.4x1016 cm-2 SHEi stress at VGB = 12 V, VCB = 10 V, and IG = InA. (c) CHEi stress at VGB=VDB= 16 V and
floating VSB and ID = 1 uA for 1 sec (curve 1) and 500 s (curve 2).















CHAPTER 5
PROFILING OF INTERFACE TRAPS
BY THE DCIV METHOD

5.1 Introduction

Due to the localized nature of channel-hot-carrier induced degradation, multidimensional transistor simulators such as PISCES or MINIMOS are usually employed to analyze the effects of channel-hot-carrier induced degradation on the electrical performance of the transistors [50, 76-83]. However, lack of reliable experimental data has often required using assumed spatial distribution of interface and oxide in transistor simulations. Hence, there is a need for a reliable characterization technique to profile the spatial distribution of stress-generated interface and oxide traps where the experimentally-determined interface and oxide trap profile can be fed into the simulators to analyze the electrical characteristics of a degraded transistor.

Two experimental techniques were recently proposed to characterize the spatial distribution of stress-generated interface traps (QIT) near the drain junction of a nMOST [84-88]. Speckbacher et al [84] used the d.c. surface generation current technique to profile the spatial distribution of QIT. In this technique, the weak reverse biased drain junction current was measured in a gated diode configuration and the gate bias is varied from accumulation to inversion to vary the surface space-charge layer near the drain junction. Due to the small reverse bias applied on the drain, the profiling




61






62



capability of this technique is confined to region very close to the drain junction. Furthermore, the sensitivity was limited by very low stress-generated currents (- 5fA) superimposed on large unstressed leakage current. Ancona et al [85], Heremans et al [86], and Ma et al [87-88] employed the charge-pumping current technique [89-92] which is a variation of the well-known pulse field-effect method [93-95]. The spatial distribution of QIT is obtained by correlating the drain voltage pulse during the charge filling (accumulation) cycle with the width of the surface space-charge layer. However, as the charge-pumping technique is based on a dynamic electron-hole capture and emission events at the interface traps, interpretation of charge-pumping results critically depends on the shape, rise and fall times, amplitude, base reference level, and frequency of the gate pulse and the drain/source biasing condition to control the electron-hole emission and capture rates [86]. Ancona et al [85] and Ma et al [88] also cautioned the use of proper pulse alignment between the gate and drain (1800 out-of-phase) to avoid erroneous interpretation of the QIT spatial distribution. In this chapter, we will show that the DCIV method can be extended to profile spatial distribution of stress-generated QIT [96].

5.2 Theory

The DCIV Profiling Technique employs the IB-IC components to determine the lateral distribution of the interface traps. The magnitude of IB-IC is the areal integration of the surface recombination rate Rss multiplied by the electronic charge q. Assuming a constant channel width W, IB-IC can be expressed as





63



rL-Ys
IB-c = q W RSS dy. (5.1) YD

In Figure 5. 1, the drain-body metallurgical n+/p junction is referenced as the origin of the y-direction. Thus, YD is the drain surface-charge layer thickness, ys is the source surface space-charge layer thickness, and L is the channel length.

In Eqn.(5. 1), the surface recombination rate Rss can be independently controlled by VGB and VBE. VGB controls the surface band-bending condition and can be related to surface band-bending potential xs using Eqn.(3.32). VBE determines the separation between the quasi-Fermi potentials for electron FN and hole Fp and is related by VBE = (FN Fp)/q. As shown in Eqns.(3.9) and (3.12), the surface band-bending and quasiFermi level splitting control the surface electron and hole concentrations. Expressing Rss = Nrr(y).R s-eff where NyT is the areal density of the interface traps and Rs_-eff is the effective recombination rate, IB-IC can be further expressed as L-Ys
IB-1C = q W J NIT(y) Rs-eff dy. (5.2) YD

It is reasonable to assume that the surface band bending (Ws) and the quasi-Fermi potentials (FN and Fp) are position-independent for small channel current and long minority carrier diffusion length (LDiff > L). Then Rs-eff can be treated as a constant and independent of position and Eqn.(5.2) becomes [L-ys
IB-1C = q W Rs-eff NIT (Y)dy. (5.3) YD





64



Collector Emitter D Drain G Gate Source S


Si02



SL 1 > Ys
I YD L YS p-Base

I IB t Base, B
Body, X


4






I -0- 0 unstressed
II

OV VCB I I I I
y= YD y L-ys L




Figure 5.1 (a) Cross-sectional view of the nMOST transistor used in the SE-DCIV
profiling configuration. (b) Base recombination current versus drain bias, Ix VDX or IB VCB, for five interface trap distributions. Curve-#(Drain, Channel, Source) are: Curve-O (0,0,0), Curve-i (nonuniform, 0, 0), Curve-2 (0, constant, 0), Curve-3 (0, nonuniform, 0), and Curve-4 (nonuniform,
nonuniform, 0).






65


From Eqn.(5.3), the decrease in IB-IC for an incremental increase in YD can be expressed as

AIB-1C (Yd)
= -q W R_-eff NT(yd) (5.4) Ay

In the limit of infinitesimal Ay, Eqn.(5.4) becomes

IB-IC (Yd)
= -q W Rs-eff NIT(yd) (5.5) ay

Then, NIT(YD) can be related to the IB-IC and VDX by the following relationship.

a IB-1C (Yd) IB-1C (Y) ay
(5.6)
aVD ay aVDX

Substituting Eqn.(5.6) into Eqn.(5.5) and rearranging.

aYD B-1C
NIT (YD) = W Rs-eff -D' I a [ (5.7) Eqn.(5.7) is the DCIV interface trap profiling equation. It shows that the interface trap concentration is proportional to the slope of the experimental IB-Ic-VDX data. To obtain the spatial distribution of interface trap along the channel, a reverse-bias voltage is applied to the drain junction to increase the drain/body surface space-charge layer. On the other hand, to probe the traps located within the drain surface space-charge layer, the drain/body junction can be slightly forward-biased. The extent of narrowing the drain/body surface space-charge layer will depend on the relative magnitude of VDX and VBE. Evidently, an alternate approach is to forward-bias the drain junction and its baseinput conductance versus the forward-bias voltage which could also give an indication






66



of the spatial distribution of the interface trap in the region near the drain's metallurgical p/n transition boundary.

To obtain a qualitative value for NIT, a model for Rss needs to be assumed. Analytical models for Rss can be a single energy-level QIT [59, 97] or a distributed DIT that is uniformly distributed in energy near the middle of the silicon energy gap [66, 73]. The conversion of VDX to y-position to give the y-dependence of NIT can be obtained using a MOST device simulator with known doping profile and device's structure. However, an analytical abrupt n+/p junction approximation [98], given below, can be used as a first order estimation to relate the drain surface space-charge layer thickness to the drain voltage.

YD(VDx) = [ (2Esi/qNAA) (VDX-bi+VDS-_ S) ]1/2 (5.8) where ESi is the silicon dielectric constant, VDX-bi = (kBT/q)loge(NDD+NAA/ni2) is the built-in potential of the drain n+/p junction at the interface. The silicon surface energy band bending in the x-direction at the drain, ps, depends on the applied gate voltage and can be determined from the one-dimensional electrostatic formula of a MOS capacitor. NDD+ is the donor concentration in the heavily-doped drain and NAA is the acceptor concentration in the p-base. Note that in the case of MOST with a lightly-doped-drain structure, the yD(VDx) estimated using Eqn.(5.8) will be larger than the actual value as the voltage drop across the lowly-doped drain region is not considered by this simple approximation.

Based on Eqn.(5.7), qualitative interpretations on the spatial distribution of the interface traps can be deduced. For example, if there is a non-uniform distribution of





67



interface traps along the channel and near to the drain region of a nMOST, the magnitude of IB-IC decreases when the electronic boundary of the surface drain/body space-charge-layer extends towards the center of the channel by a positive VDX. The decrease in IB-IC is caused by the decreased number of interface traps available for recombination in the portion of channel modulated by VDX. Figure 5.1(b) shows the five variations of the IB current with VDX for five possible NIT spatial distributions. Curve (0) is the IB characteristic for a fresh unstressed nMOST. For low and uniform NIT distribution, IB decreases with a small constant slope when VDX increases. Curve

(1) shows a non-uniform NIT distribution within the drain/body surface space-chargelayer, i.e. 0 < y < yD, and uniform NIT distribution along the channel. Curve (2) exhibit a constant AIB/gVDX slope and hence indicates a uniform NIT distribution along the channel. For VDX < 0, the constant IB also suggests uniform NIT distribution in 0 < y < yD. Curve (3) shows that IB decreases with a varying slope for VDX > 0 V and thus indicates non-uniform NIT distribution. A uniform NIT distribution is implied in 0 < y < yD by using the same interpretations given above. Curve (4) illustrates the case for a non-uniform NIT distribution along the channel and also within the drain surface spacecharge-layer.

5.3 Experiment and Results

Industrial n-channel MOST is used in the demonstration of this DCIV Profiling technique. The n-MOST is fabricated in a p-well with n-epilayer/p-substrate and has drawn channel length L=1.6pm, drawn channel width W=100.01tm, gate oxide thickness xox=160A, and p-well doping concentration NAA=l.4x1016cm-3. The SE-





68



DCIV configuration is used to inject minority carriers (electrons) into the p-body from the forward-biased source junction. In this illustration, we generate non-uniform NIT by stressing the nMOST under CHE stress condition (VDs=8.0 V, VGS=4.0 V, and Vsx=0.0 V) for 120 minutes.

Figure 5.2 shows the post- and pre-stressed DE- and SE-DCIV characteristics. Both DE- and SE-DCIV are measured at VBE=0.4V and VCB=0.OV. For the case of SEDCIV measurement, the post-stressed IB increase for -IV < VGB < OV is attributed to stress-generated interface traps along the channel which results in more electron-hole recombination through these additional interface traps in the channel. New bulk defects are not created by the CHE stress; otherwise, the measured post-stressed IB will show a relative constant increase over the un-stressed IB-VGB data because the bulk diffusionrecombination Ig components (IB-b2 and IB-bl in Figure 3.2) are independent of surface modulation by the applied VGB. There is also negligible shift of the post-stressed IB peak, in comparison with the unstressed Ig peak at VGB=-0.35V, along the voltage-axis. This suggests that there is negligible built-up of oxide charge during the 120-minutes CHE stress. Since the channel hot electrons experience the largest accelerating electric field near the drain/body metallurgical junction, we anticipate that there should also be a spatial distribution of stress-generated interface within the surface drain space-chargelayer. Thus the drain surface-charge-layer recombination current IB-2S, as discussed in chapter 3, should dominate when the drain is used as an emitter in the DCIV measurement. Indeed, this was observed as indicated by post-stressed DE-IB-VGB in Figure 5.2.





69









10-8 I I I I I I I Top-Emitter BJT
DE Drain-Emitter, Source=Collector SD Source-Emitter, Drain =Collector

10-9_ stressed (DE).



01 stressed (SE)

10-10
-- 1 _-unstressed



10-11,, 1 1
-2 -1 0

VGB/(1 V)



Figure 5.2 Effect of channel hot electron stress on the Drain-Emitter and SourceEmitter DCIV characteristics. All measured at VBE= 0.4 V, VCB= 0.0 V and 297K. CHE stress conditions are VDS= 8.0 V and VGS= 4.0 V for 120 minutes. The solid line refer to the post-stressed measurements while the dotted line is for the unstressed measurement. The unstressed IB-VGB measurement is similar for both drain and source junctions and hence only
one curve is shown.





70



Figures 5.3(a) and 5.3(b) shows the AIB-VBE curves at three different gate biases (VGB=-0.5V, -1.OV, and -2.OV) for SE- and DE-IB-VBE measurement configuration, respectively. In Figure 5.3(a), the increase in AIB for VGB=-0.5V is primarily contributed by the surface recombination along the channel, i.e., IB-IC component. The surface band-bending potential Ns at VGB=-0.5V is computed to be equal to 0.08V, which is close to the flat-band condition (VGB=-0.65V) obtained by a high-frequency Cg-VG measurement on a 450p.mx450pm capacitor. At low VBE bias, it is observed the ideality factor n is about 1.0. This suggests that the hole surface concentration Ps is greater than the surface electron concentration Ns which results in the surface recombination rate to be dependent on VBE. Thus AIB oc IB-iC exp(qVBE/kBT). At high VBE bias, it is observed that there is a change of the ideality factor from n=1 to n=2. This is consistent with Rss o exp(qVBE/2kBT) for Ps = Ns, as discussed in Chapter 3.2. The transition from n=1 to n=2 occurs at about 0.55V and agrees reasonably well with the theoretical value predicted using a simple single energy-level interface trap model.

VBE = 2 [VF rS (VGB--05V) ] (5.9)
= 2[(kBT/q) log(NA/ni) 0.08V]

= 2[0.36V 0.08V]

= 0.56V

In Figure 5.3(b), the increase in AIB for three gate bias values is attributed to the surface recombination within the drain/body surface space-charge-layer. This is supported by





71





10-6

10-7 Source-Emitter BJT

10-8
VGB=-.5V
10-9
C 0-0--1.0V
-2.OV
10-11 /

10-12 -- (a)

10-6
Drain-Emitter BJT 10-7
A VGB=-0.5V 10-8 -1.0V I -2.OV

10-9 n=1.7

10-10



10-12 I- (b)
0.2 0.3 0.4 0.5 0.6 0.7 VBE/(1 V)


Figure 5.3 Effect of VGB on the AIB-VBE characteristic for (a) Source-Emitter BJT and
(b) for Drain-Emitter BJT. All measured at VCB= 0 V and 297K.





72



the experimentally observed n= 1.7 ideality factor obtained by fitting AIB-DE-VBE to a single exponential function [AIB=AgBexp(qVBE/nkBT)].

Figures 5.4(a) and 5.4(b) show the post- and pre-stressed Source-Emitter IB-VCB experimental results, AIB-VcB and -[a(AIB)/aVcB], respectively. In Figure 5.4(a), the unstressed Ig is essentially flat for -0.2V < VCB < 2.5V which suggests low and uniform interface trap density. This is intuitively expected for an unstressed MOST fabricated by the state-of-the-art MOS technology. The sharp rise in In for VCB <

-0.4V is due to the additional base current contribution by the slightly forward-biased drain-body junction. This excess base current has an exponential dependence on the magnitude of VCB which will overwhelms the background base current. The slope of the post-stressed IB data is observed to change at different value of VCB. This suggests a non-uniform distribution of stress-generated NIT. The downward trend of both the postand pre-stressed base current in Figure 5.4(a) when VCB > 3.0OV indicates hole current from interband impact generation by the injected electrons which are accelerated by the high electric field in the revere-biased (-3V) drain junction space charge-layer. Thus, this limits the range of VCB that can be applied to extend the drain/body space-chargelayer out into the channel without causing any additional stress on the device during profiling measurement. To extract the stress-generated NIT from the measured post- and pre-stressed IB-VCB, we perform the following analysis on the data. The difference result is shown in Figure 5.4(b).

AIB = [ IB (VBE=0.4V, tstr=120mins) -IB (VBE=0. 4V, tstr=0) ]

[ IB (VBE= OV, tstr,=120mins) -I (VBE= OV, tstr=) ] (5.9)
= AIB-1C





73



300

250

200
I stressed 150


(a)
50
30 120

-stressed 110


-100 0

> 90 m


80

0o .70 (b)
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 '- VCB / (1V)
+ 0.1 CHANNEL y / (m)
( I I I I I (c)
0 0.2 0.3 0.4 0.5 0.6



Figure 5.4 Stress-and-measure (SAM) data of base recombination current variation
with drain bias using forward-biased n+ source/p-body junction as the minority carrier (electron) emitter. CHE stress at VDS= 8.0 V, VGS= 4.0V and Vsx= 0.0 V for 120 minutes. Measured at VBE= 0.4 V, VGB= -0.5 V, and VCB= 0.0 V. (a) Raw data showing decrease at VDX >-3 V due to hole generated by electron impact. (b) CHE generated excess base current, AIB, and a(AIB)/aVcB versus VCB. (c) Estimated interface trap position from the
n+ drain junction boundary. (lpAN corresponds to NIT = 109 traps/cm2.)





74



In Eqn.(5.9), the first subtraction used is to eliminate the contribution from the drain surface recombination IB-2S component. The second subtraction used is to cancel out the impact-generated hole current component. The final subtraction gives the stressgenerated base current, AIB, as shown in Figure 5.4(b). It is observed that there is a continuous rise of the IB (<-77pA) from the source towards the drain (-117pA) which is expected from the channel hot electron stress. This AIB-VCB data is then least-squarefitted to obtain the slope of a(AIB)/aVCB which is shown in Figure 5.4(b). Evidently, there is a change in slope in Ig as VCB varies which indicates a non-uniform distribution of stress-generated interface traps. The results qualitatively shows that the stressgenerated interface trap starts to increase at about 0.6jim from the n+ drain junction boundary( a(AIB)/aVcB 6pA/V), reaching about four times (-24pA/V) at 0.2gim from the n+ drain junction boundary. This is consistent with the spatial variation of the electric field where electric field strength increases after the depletion point, y = y(VDsat), and reaches a maximum at the drain metallurgical junction.

In summary, a simple and reliable extension of the DCIV technique is demonstrated to profile the stress-generated interface trap densities along the channel of MOS transistors.














CHAPTER 6
LINEAR REDUCTION OF DRAIN CURRENT WITH
INCREASING INTERFACE RECOMBINATION IN NMOS
TRANSISTORS STRESSED BY CHANNEL HOT ELECTRON

6.1 Introduction

Recently, Neugroschel and Sah proposed a simple and highly sensitive method, the Direct-Current Current-Voltage (DCIV) method, to monitor the degradation of silicon MOS transistors (MOSTs) under channel and substrate hot carrier stresses [41]. The DCIV method uses the gate-voltage-controlled interfacial recombination current IB (base or body current) to monitor the generation rate of interface traps and the charging rate of oxide traps in MOSTs. However, the traditional transistor reliability measurements monitor the changes of the saturation drain current (AID-sat), threshold gate voltage (AVGT) and subthreshold voltage swing (AS) [33-37, 51-53, 98-99]. In this chapter, we demonstrate the anticipated linear dependencies of these traditional characteristics on the DCIV base-body recombination current, i.e. AIDsat, AVGT, AS o AIB.

6.2 Experiments

The n-channel MOSTs (nMOSTs) used have a drawn channel width/length aspect ratio of L/W = 17.5g.m/0.35pm, a 60A gate oxide, and separate source, drain, gate, body contact pads. During the Stress-and-Measure (SAM) experiment runs, the nMOSTs were stressed by channel hot electrons (CHEs) in the forward configuration (i.e.




75






76


n+drain/p-body junction is reverse-biased) while their electrical characteristics were measured after each stress duration in the reverse configuration (i.e. n+source/p-body junction is reverse-biased to measure AIs-sat, AVGT and AS, and the n+drain/p-body is forward-biased to measure AIB in the DCIV mode which is known as the Drain-Emitter DCIV or DE-DCIV ). This interchange of source and drain during measurements is traditionally employed because the generated interface traps and charged oxide traps are located over the drain-junction space-charge region. Two bias voltage conditions are used during the CHE stress duration: (i) the traditional VDS = 2VGs which gives the maximum body current, Ix-stress, and (ii) the constant VDS and VGS with VDs>>VGS-VGT to probe a wide range of hot electron kinetic energies. In each SAM experiment, the drain saturation current is measured at VDS=VGs= 1.0 V and Vsx= 0.0 V. The threshold voltage is defined as the gate voltage at ID = 500 tA with VDS= 0.5 V and Vsx= 0.0 V. The subthreshold slope is computed at ID = 100 nA with VDS= 0.5 V and Vsx= 0.0 V. The n+drain/p-body junction is forward biased at 0.4 V and the n+source/p-body is short-circuited for the IB-VGB measurements in the DE-DCIV mode.

6.3 Results and Discussion

Figure 6.1 shows typical IB-VGB DE-DCIV characteristics with increasing accumulated stress time of a nMOST stressed by CHEs. The stressed voltages were VDS= 3.3 V and VGS= 1.6 V, corresponding to peak body current condition. The IB-peak increases with CHE stress time and occurs at a nearly constant VGB = -0.45 V. The rise of IB-peak is due to the stress-generated SiO2/Si interface trap over the drain junction space-charge region [84-88, 96]. The negligible VG shift of the InBpeak indicates






77






150
DE-DCIV
2.0x104s VDX=-0.4V

100
0 556s

SOs

m 50




0 1I I
-1.0 -0.5 0.0 0.5

VGB/(1 V)


Figure 6.1 Drain-Emitter DE-DCIV characteristics of a nMOST stressed by channel
hot electrons.






78


negligible charging of the oxide traps over the drain-junction space-charge region because the CHEs do not have sufficient kinetic energy, KE = q(VDs-VDS-sat) = q[VDs-(VG-VGT)] = 3.3-1.6 = 1.7 eV, to surmount the 3.13 eV SiO2/Si barrier. The CHEs also do not have sufficient kinetic to create secondary hot holes, via interbandimpact electron-hole pair generation, to surmount the 4.25 eV Si02/Si hole barrier, which would require a minimum energy of q(VDs-VDS-sat) = 0B-hole Egap-Si = (4.25 1.12) eV = 3.13 eV via the Auger generation pathway [117, 23]. The forward and reverse IDsat and ID-VGs characteristics of the nMOST is also shown in Figures 6.2 and 6.3, respectively. The asymmetrical difference between the post-stressed forward and reverse characteristics clearly indicate the localized CHE-induced degradation near the drain junction of the nMOST.

Figure 6.4 shows the linear dependence of Als-sat and AVGT on AIB-peak for the two sets of stress voltage conditions. This linear dependence is anticipated from the simple textbook theory of ID-sat and VGT. Carrier mobility degradation in the channel should be insignificant since the stress-generated traps are localized at the drain-junction. From the simple parabolic MOST equation [99],

ID-sat = (W/2L)COIR(VGs VGT) 2, (6.1) we get

AIDsat/IDsat0 = (2AVGT) / (VGs-VGTO) (6.2) where the subscript 0 denotes the pre-stress value. Eqn.(6.2) assumes that (VG-VGT) >> AVGT/2. Since AVGT = q(ANOT + ANIT)/Cox and IB O NIT [59, 99-100], we can immediately obtain AID-sat oc AVGT AIB.






79







1.5

F=Forward R=Reverse ,,. F



CO
0
Pre-stress .-c S 0.5 Post-stress





0.0 I I I I
0.7 0.8 0.9 1.0 1.1 1.2

VD=VG/(1 V)


Figure 6.2 Forward and reverse IDsat characteristic of a nMOST stressed by channel
hot electrons. Difference between the post-stress forward (F) and reverse (R) characteristics reveal the asymmetrical nature of channel hot electron stress. tstress= 20 ksec. Pre-stress forward and reverse IDsat characteristics
are identical.





80









10-4 F=Forward R=Reverse
10-s R

10-6 Pre-stress F

10-7 Post-stress

10-8



10-10

1011
0.0 0.2 0.4 0.6 0.8 1.0 VGs/(1 V)


Figure 6.3 Forward and reverse ID-VG characteristic of a nMOST stressed by channel
hot electrons. Difference between the post-stress forward (F) and reverse (R) characteristics reveal the asymmetrical nature of channel hot electron stress. tstress= 20 ksec. For clarity, only the unstress forward ID-VG
characteristic is shown. All measured at VDS= 0.5 V.





81







10- 100




10-2 .V t --VG 10'0AS


) C :




S 00 0
10-3 D do 10-3



10-4 I I I 10-4
1001 10 10 103 %AIB-DE-peak



Figure 6.4 Dependence of -AID-sat/'D-sat, AVGT and AS on AIB.peak/IBpeakO in 0.35gm
silicon nMOSTs stressed by channel hot electrons.
A,O,0: VDS/VG=4.0/1.9, 3.75/1.8, 3.3/1.6.
A,O,IM,: VDS=4.OV, VGS= 1.0, 2.0, 3.0, 4.OV






82


The stress-induced widening of the subthreshold swing, AS, is also shown in Figure 6.4. It increases initially as AS o AIB, then rises faster, and finally approaches again AS Oc AIB. This is consistent with the simple subthreshold slope expression,

S = 2.303 (kT/q) [1+ (Cit+Cs)/Cox] (6.3) where Cit = qDIT and NIT = I D IT(ET)fT(ET)dET [100], which indicates an initial and final linear dependence, AS oc ACit oc ANT OC AIB, while the superlinear rise of AS in the mid-range of stress is caused by the faster change of the surface potential with AVGT relative to the energy distribution of DIT(ET).

In summary, an experimental proof is given of the theoretically anticipated linear dependence of AIB-peak on the traditionally monitored transistor reliability characteristics, AID-sat and AVGT. This correlation provides the basis for using the DCIV method to monitor MOS transistor degradation.















CHAPTER 7
INTERFACE TRAPS GENERATION MODEL

7.1 Introduction

As shown in Chapters 5 and 6, interface traps are generated when MOS transistors are stressed under channel hot carrier stress conditions. These interface traps are electron-hole recombination-generation sites and thought to be dangling silicon and oxygen bonds (Si., SiO*) [101] created by energetic electrons and holes breaking the weak intrinsic (Si:Si, SiO:Si) or impurity (Si:H, SiO:H) bonds. However, the physical mechanism of interface trap generation during channel hot carrier stress was still not well understood. No agreement currently exists on this point in the literature. Some studies attributed interface trap generation to electron injection [35, 102-103] where a critical energy of 3.5 eV is needed for the electrons to be injected into the oxide to break the Si:H bond [35]. Others propose that hot holes are the responsible carriers [104-106]. There was also some independent evidence that recombination of electrons with trapped holes in the oxide also generate interface traps [7-8]. It was also proposed that the energy released by electron-hole recombination in the Si02 dissociates a weak Si:H bond, thus the generation of interface traps involve hot electrons, hot holes, and hydrogen [106].

In this chapter, a new interface trap generation model, formulated by Sah [42], is presented to account for the channel-hot-carrier-induced degradation. This model is



83






84



based on a physical bond-breaking mechanism and considers the bond-breaking efficiency of the hot electrons and holes along the surface channel on oxidized silicon. It identifies the primary and dominant hot carriers (electrons or holes) that are involved in the interface trap generation process. In this chapter, the analytical interface trap generation theory is presented and the validity of the model is supplemented by experimental verification.

7.2 Theory

Consider a n-channel MOS transistor biased in the flat-band condition shown in Figure 7. l1(a), the drain n+/p junction is reverse biased by VR which results in a wider space-charge-layer (SCL). The corresponding energy band diagram is depicted in Figure 7.1 (b). The drain junction space-charge-layer can accelerate electrons, injected at the low-field edge (LFE) of the drain SCL, and holes, injected at the high-field edge (HFE) of the drain SCL, to higher kinetic energy when the electrons or holes transit across the SCL. When the accelerating electrons and holes acquire sufficient energy, greater than the bond-breaking energy (qVBB), they can break the hydrogenated interfacial bonds (Si:H and SiO:H) to release the hydrogen and create new interface traps.

Based on Sah's interface trap generation model, the rate of interface trap generation (dNIT/dt) is proportional to the injected carrier flux density (Jt/q), the probability of surviving a scattering, and the probability of bond-breaking. Thus, the interface trap generation rate is given by the product of these three terms:






85






SiO2 O NIT AED
(LFEi)
(HFHi) ---t- (EIH) (STH) L.
n+Si (ITHi), O, : n+Si
COLLECTOR I I I EMITTER
SOURCE I I DRAIN

Y=Ysc

y=O \
p-Si
BASE



B I1 (a) y=0 Y=Ysc
H- Ysc-y.
y=daselec
-V (LFEi) d BBelec- -el
--F Ic VFN+ (LF.. C,S Ec A A
ET F V B-ele
Ev 96*0303CTS V
Vb VF VR + Vu
VFN+


V-oe (Hi): VFNTi IE +VR
(IT H i) E E ,
dsBB-hole AEc ED E VEI-hase EV (HFHi) (EIH) (STH)
dBB-hole (EIH) (STH) (b)




Figure 7.1 Cross-sectional view of a reverse biased p/n+ junction covered by SiO2
showing three dominant bond-breaking pathways to generate interface traps. (a) Cross-sectional view. (b) Transition energy band diagram along the oxide/silicon interface. Bond breaking is marked star and vertical dash line. Open triangles are interface traps. Filled triangles are Si:H and SiO:H. Dots are electrons and circles are holes. Electron and hole motions are marked by arrows with solid and open tips, respectively. Adapted from
Sah [42, p.123].






86



dNT -dBB j Ysc r (y-dBB) dy
= exp -exp (7.1)
dt q 2 dBB A3 ABB

The product of the first and second term in front of the integrand of Eqn.(7. 1) gives the decrease of the ballistic hot carrier density due to scattering prior to breaking a bond. The first term inside the integrand of Eqn.(7. 1), exp[-(y-dBB)/X3], is the fraction of the injected carrier which has escaped scattering and not broken bond when the carrier transits from y=(y-dBB) to y=y. (dy/?BB) is the fraction of bond-breaking events within the dy region. The combined mean-free-path, X3, contains three components: the surface scattering, Xs, which includes the phonon, impurity, electron-electron and electron-hole scattering events; the interband impact-generation of electron-hole pair, pN; and the bond-breaking by the energetic carrier, BB.

1/A3 = 1/As + l/ApN + 1/ABB (7.2)

= 1/A2 + 1/BB (7.3) Since the bond-breaking event occurs infrequently, it is anticipated that gBB >> XPN > 'S > X2 > X3 Solving Eqn.(7. 1) and using the approximation (Ysc-dBB)>>3,

dNIT JI -dBB 3 (Ysc-dBB)
- exp J 1-exp (7.4)
dt q A2 BB 3 BI d AA
= exp- 0 (7.5) q 2 BB

The interface trap generation efficiency, 1iT, is defined as the rate of interface trap generation per injected carrier density. Hence, Eqn.(7.5) can be expressed as follows.






87



IT = NT / [- (7.6) dt q


= exp (7.7)
2 BB


IT0 (7.8) BB

where TiTO = exp(-dBB/X2) is the normalized interface trap generation efficiency. Analytical expressions for voltage dependence of the bond-breaking distance, dBB, can be derived by using the depletion approximation of an abrupt n+/p junction and by tracing the three bond-breaking pathways, illustrated in Figure 7.1, of the energetic electrons and holes.

First, let consider the electron injection from the low-field edge of the drain spacecharge layer (the LFEi pathway shown in Figure 7.1), the accelerating electron in the drain space-charge layer can break an interfacial bond when it has kinetic energy greater than the bond-breaking energy, qVBB. Thus, dBB can be expressed as follows:


dBB = (2/qNAA) ( VBB ) (7.9) where s is the silicon dielectric constant, q is the electronic charge, and NAA is the acceptor doping concentration. For this LFEi interface trap generation pathway, the bond-breaking distance is independent of the applied voltage and only depends on the acceptor doping concentration, NAA, and the threshold bond-breaking energy, qVBB. This implies that the efficiency of interface trap generation by these energetic hot electrons is a constant when the applied voltage exceeds the bond-breaking energy, and






88



is zero if the applied voltage is below. The normalized interface trap generation efficiency for this LFEi pathway is.


17ITO = exp[- (2 /qNA) /A2] VBB (7.10)

Next, consider the hole injection pathway from the high-field drain edge labelled HFEi in Figure 7.1 (a). The sources of these holes can be due to thermal generation or from the interband-impact generation process initiated by the energetic electron when the electron kinetic energy, KEelectron, is greater than the electron-hole pair generation threshold (EPN > EG-Si 1.17eV) [109]. Typically, the concentration of the ElectronImpact-Generated (EIH) holes is higher than the thermally-generated hole concentration. These impact-generated holes can be back-injected into the drain spacecharge layer from the maximum-field edge and break the interfacial bonds when its energy exceeds the bond-breaking energy. The bond-breaking distance for this HFHi interface trap generation pathway is:


dBB = (2E/qNA) [VR+VBI VR+VBI-VBB ] (7.11) Similarly, the corresponding normalized interface trap generation efficiency is given by.


7ITO = exp[- (2e/qN ) /A] VR+VBI VR+VBI-VBB ] (7.12) where VBI is the built-in potential of the abrupt n+/p drain junction. Comparing with Eqns.(7.9) and (7.10), eqns.(7.11) and (7.12) reveal that this HFHi interface trap generation pathway has a very strong dependence on the applied stress voltage, VR. This characteristic strong voltage dependence of dBB or rrtTO provides a feature for the theory to be experimentally verified. This will be described and elaborate in the






89



subsequent section. Eqns.(7.11) and (7.12) apply only to the drain-current range of the MOST where VR=IVD VDsatl > 0 where VDsat is the drain voltage at drain current saturation and may approximated by IVG VGTI. The VBI must also be modified to take into account of the built-in potential barrier height difference between the drain/body and source/body junctions. For identical drain and source junction, VBI=0 when the MOST's channel is inverted.

Another possible pathway for hole to generate interface traps is via the Interband Tunneling Hole Injection (labeled as ITHi in Figure 7.1). This ITHi pathway will become important in heavily doped n++emitter/p+base junction of advanced submicron BJTs and future generation of MOSTs where the body doping concentration is approaching and exceeding 1018 cm-3. At this high concentration, valence electrons in the body(base) region can quantum mechanically tunnel into the empty states of the conduction band in the quasi-neutral drain (emitter) region, leaving behind thermal holes. Similarly to the interband impact-generated holes, these interband tunneling holes can then accelerate in the high-field SCL to create new interface traps. The bondbreaking distance associated with this ITHi interface trap generation pathway is given as:


dBB = (2E/qNAA) [VR+VBI-VSi-Gap-VFN+


VR+VBI-VSi-Gap-VFN+- VBB ] (7.13) In Eqn.(7. 10), the valence-electron tunneling process decreases the voltage dependence of the bond breaking distance by (VsiGap + VFN+) where VSi-Gap is the silicon energy






90



gap potential and VFN+ is the Fermi level potential above the conduction band edge of the quasi-neutral n++ drain-emitter. The corresponding normalized interface trap generation efficiency is then given by:


ITO = exp[- (2/qN,) /A2] [ VR+VBI-VSi-Gap-VFN+


VR+VBI-VSi-Gap-VFN+- VBB ] (7.14)

As shown in Eqns. (7.12) and (7.14), the interface trap generation efficiency has a strong voltage dependence on the applied stress voltage when the electron (or hole) enters the space-charge-layer from the high-field edge. On the contrary, if the electron (or hole) enters from the low-field edge of the space-charge-layer, the interface trap generation efficiency has no voltage dependence, as seen from Eqn.(7. 10). Hence, a voltage-dependency experiment can be designed to correlate the experimental results with the interface trap generation theory presented in this section.

7.3 Experiment and Results

Four sets of Time-to-Failure SAM experiments on silicon nMOST, pMOST and npnBJT were analyzed using the new interface trap generation theory. The nMOST data are obtained from a channel-hot-electron (CHE) stress experiment where the nMOST test structure was fabricated from a 1-micron BiCMOS technology with 100 Ltm channel length, 100 lm channel width, and 160 A thick gate oxide. The pMOST data are similarly obtained from a channel-hot-hole (CHH) stress experiment where the pMOST test structure was fabricated from a 0.8-micron CMOS technology with 20 Lm channel length, 20 jim channel width, and 110 A thick gate oxide. The experimental






91


details of the nMOST and pMOST SAM data will be described in Chapter 8. The two sets of npnBJT data are obtained from emitter-base reverse-bias stress experiment. The npnBJTs were fabricated from a submicron BiCMOS technology with 7x 1017 cm-3 base doping concentration and lOmx0.6pm emitter contact area. During the emitter-base stress experiment, the emitter-base junction of the npnBJTs was reverse-biased at VR while the collector-base junction was either open-circuited (OC) or forward-biased

(FC). A 10% change in the common emitter d.c. beta gain was used as a failure criteria in the TTF SAM experiments.

Figure 7.2 plots the experimental and theoretical interface trap generation efficiency for the nMOST, pMOST and npnBJTs versus the applied stress voltage. The interface trap generation efficiency was obtained by taking the inverse of the TTF SAM data and translating the efficiency data vertically by normalizing for the difference in stress current density and contact area in each SAM experiment. The applied stress voltage is the magnitude of the drain voltage for the nMOST and pMOST data and the magnitude of reverse-biased emitter-base voltage for the npnBJTs. The nMOST and pMOST experimental data were least-square-fitted using Eqn.(7.12) where VBI=O and VR=VD-VDsat. The LSF results give bond-breaking energies of qVBB= 3.087 eV and standard deviation of 1.12 % for nMOST, and 3.040 eV and 1.67 % for pMOST. The HFHi npnBJT data (inverted triangle in Figure 7.2) was also least-square-fitted using Eqn.(7.12). The LSF results gives a bond-breaking energy of qVBB= 1.876 eV and a standard deviation of 1.51%. A reasonable built-in potential of VBI= 1.11 V was also obtained from the LSF, reflecting the highly doped base region in advanced submicron






92










10-7 i I I I 960711MH 90303CT
10I
npnBJT 1 ox.6g2 *"' 20x1.272
10-9 HFHi .-- pMOST
J \ ,-6 HFEi
S10- 1

p 1 3 npnBJT nMOST 1OOx21210-Hi HFHi 10x0o.6 2 / VGS- .=.9V
10-15 10 year TTF nMOST
1x10s LFEi
10-17 ' I
1 2 3 4 5 6 7 8
I VR /(1V)


Figure 7.2 Comparison of experimental and theoretical bond-breaking efficiency by
hot electrons and holes in Si npn BJT, nMOST and pMOST versus applied junction voltage. 10-year TTF is the TTF at the applied junction voltage indicated by the vertical arrow extrapolated by the theory from the TTF data at the higher VR for 100% or 20% AIB. Adapted from Sah [42, p.126]




Full Text
106
failure (FTF labeled on the right vertical axis). At each stress interval, the FTF is
computed using:
FTF = q-1 Jj
(8.1)
where JcH-eiectron=^CH-eiectron^drain anc* the drain contact area was used. The operation
TTF (on the left y-axis) is then obtained from
TTF0p = q-FTF(VEB)/JCH_electron(VEB=0) .
(8.2)
The validity of the current acceleration methodology is confirmed for both nMOST and
pMOST by the data overlap covering many decades of the non-accelerated TTFs (open
circles) with the accelerated TTFs (asterisk, solid triangle and circles) in the high
voltage range (5.0 V ^IVDS.stressl< 7.0 V). It also support the anticipation that a small
increase in carrier density and current in the channel from a forward-biased junction
would not significantly alter the spatial variation of the electric field, hot electron and
hole densities and their kinetic energies.
The two solid curves in Figure 8.4 were obtained from a least-square-fit (LSF) of
the eleven nMOST data points and nine pMOST data points to the theory proposed by
Sah (Eqn. (7.12)) [42] to account for interface trap generation during channel hot carrier
stress. Using a failure criteria of AIB/IB= 100% for nMOST and AIB/IB= 20% for
pMOST, the theoretically extrapolated magnitude of the maximum VDS voltages to give
TTF0p= 10 years are 4.32 V and 4.05 V, respectively, for these two technologies. An
important confirmation of the current-acceleration methodology and also the bond
breaking theory is that the LSFs of the nMOST and pMOST to theory presented in
Chapter 7 gave nearly identical hydrogen-bond-breaking threshold energy, VBB= 3.087


128
Figure 9.3. Figure 9.4 shows the extracted VGB.peak values plotted as a function of (VD
- VDsat) which is the acceleration potential through the drain-junction space-charge
layer. VDsat is the potential across the inverted electron channel and is experimentally
determined from the drain conductance measurement. For the lOOpmx 100pm
transistor, VDsat= 0.3314 V was obtained by linear extrapolating the gD data to gD= 0
A/V shown in Figure 9.5.
The dearth of data shown in Figure 9.5 could not clearly delineate the presence of a
structure near VD VDsat = 5.4 V which would correspond to the electron-impact
threshold (transition (4) in Figure 9.1) although there is a hint from the two dashed
curves in Figure 9.4. Thus a similar SAM run was performed on 20 pmx 20 pm
nMOST samples. Data were collected and analyzed following the same procedure
described for the 100 pmx 100 pm nMOST SAM experiment. For the 20 pmx 20 pm
SAM experiment, the gate voltage applied during CHE stress is common at VGS=0.8V
with additional transistors stressed near VD VDsat = 5.4 V. A 0.35 V forward-bias is
used during the DE-DCIV measurements. The drain conductance measurement on the
20 pmx20 pm transistor gives VDS.sat= 0.2174 V. Figure 9.6 shows the variation of
Voe-peak as a function of hot-electron fluence for nine VDS stress voltages. Figure 9.7
plots the VQg.p^ values, extracted at Ninj= 20x1 023 cm-2, as a function of (VD-VDsat).
The results shown in Figures 9.4 and 9.7 were combined into Figure 9.8.
Reasonable agreement between the L= 100 pm and L= 20 pm results is observed. The
intersection of the two dashed lines, obtained via two separate linear-LSF of data in the
high and low voltage range, gives VD ~VDsat= 5.479 V which agrees with the theoretical


65
From Eqn.(5.3), the decrease in IB.1C for an incremental increase in yD can be expressed
as
A^b-ic (Yd)
Ay
= -q W Rs_eff Nit (yd) .
(5.4)
In the limit of infinitesimal Ay, Eqn.(5.4) becomes
d-*-B-lc (Yd)
dy
= -q W Rs_eff NIT(yd) .
(5.5)
Then, NIT(yD) can be related to the IB_1C and VDX by the following relationship.
IB-iC(Yd) ^iB-ic (y) aY
(5.6)
dVDX dy dVDX
Substituting Eqn.(5.6) into Eqn.(5.5) and rearranging.
Njt (Yd) -
S W RS-eff
dyE
dVr
-1
ai
B1C
dVr
(5.7)
'DX u v DX
Eqn.(5.7) is the DCIV interface trap profiling equation. It shows that the interface trap
concentration is proportional to the slope of the experimental IB.lc-VDX data. To obtain
the spatial distribution of interface trap along the channel, a reverse-bias voltage is
applied to the drain junction to increase the drain/body surface space-charge layer. On
the other hand, to probe the traps located within the drain surface space-charge layer, the
drain/body junction can be slightly forward-biased. The extent of narrowing the
drain/body surface space-charge layer will depend on the relative magnitude of VDX and
VBE. Evidently, an alternate approach is to forward-bias the drain junction and its base-
input conductance versus the forward-bias voltage which could also give an indication


57
interface traps in the Si energy gap are independent of energy. In Figure 4.4(a), the
Ig.peak for the top-Emitter curve stressed at an electron fluence of 5xlOl8cm'2 occurs at a
surface band-bending of Vs = VF 0.24V and gives AS0 = 1600cm/s. The IB.peak for
the bottom-Emitter curve stressed at an electron fluence of lxl017cm'2 occurring at \\is
= VF 0.26V gives AS0 = 40cm/sec. In the above example, \j/s fr the top-E and
bottom-E curves are determined using the relations presented in chapter 3.
In Figure 4.4(b), the pre-stressed IG-VGB curve is observed to be nearly flat in the
accumulation range and sharply increases when VGB > VFB = -0.55V but subsequently
saturates to a flat plateau when VGB > VTH = +0.05V. These features can be adequately
accounted for by considering the relative geometrical increase in the emitter or collector
area contributed from the electron channel. In the accumulation range, Ic is only
proportional to the area of the n+drain/p-base well junction and thus does not depend on
Vgb. When VGB > VFB, an electron surface channel begins to form and increases the
effective collector area that collects the injected electron from the bottom-emitter,
resulting in an increase of Ic. When VGB VTH> the p-Si surface is inverted by the
gate-induced electron channel. Thus, Ic will be proportional to the sum of the area of
the n+drain/p-base well junction and this inverted electron channel. The three
characteristic Si surface potentials or Si energy band bending (FB = flatband at \j/s=0V,
INV = inversion at equal electron-hole surface concentration Ns = Ps or \j/s = VF -
Vbe/2, and TH = threshold or strong inversion at Ns = PBase or \\is = 2VF VBE) are
indicated on the pre-stress IG-VGB curve- As seen from Figure 4.4(a), the SHEi has
generated substantial interface traps indicated by the large increase in IB. But, the post-


ID/(1 A)
80
VGS/(1 V)
Figure 6.3 Forward and reverse ID-VG characteristic of a nMOST stressed by channel
hot electrons. Difference between the post-stress forward (F) and reverse
(R) characteristics reveal the asymmetrical nature of channel hot electron
stress. tstress= 20 ksec. For clarity, only the unstress forward ID-VG
characteristic is shown. All measured at VDS= 0.5 V.


5 PROFILING OF INTERFACE TRAPS BY THE DCIV METHOD 61
5.1 Introduction 61
5.2 Theory 62
5.3 Experiment and Results 67
6 LINEAR REDUCTION OF DRAIN CURRENT WITH INCREASING
INTERFACE RECOMBINATION IN NMOS TRANSISTORS
STRESSED BY CHANNEL HOT ELECTRONS 75
6.1 Introduction 75
6.2 Experiments 75
6.3 Results and Discussion 76
7 INTERFACE TRAPS GENERATION MODEL 83
7.1 Introduction 83
7.2 Theory 84
7.3 Experiments and Results 90
8 PHYSICS-BASED TIME-TO-FAILURE EXTRAPOLATION
ALGORITHM USING CURRENT-ACCELERATED CHANNEL-
HOT-CARRIER STRESS 95
8.1 Introduction 95
8.2 Methodology of Current-Accelerated Channel-Hot-Carrier
Stress 96
8.3 Results and Discussions 100
9 HOT HOLE INJECTION INTO Si02 IN N-CHANNEL MOS
TRANSISTOR DURING CHANNEL-HOT-ELECTRON STRESS 117
9.1 Introduction 117
9.2 Four Fundamental Interband Hot Hole Generation Pathways 119
9.3 Experiments 122
9.4 Results and Discussions 123
10 REDUCTION OF INTERFACE TRAPS IN P-CHANNEL MOS
TRANSISTOR DURING CHANNEL-HOT-HOLE STRESS 138
10.1 Introduction 138
10.2 Experimental Results 138
10.3 Kinetic Model 143
10.4 Summary 147
IV


46
the IB peak occurs when the surface electron and hole densities are comparable to each
other.
3.5 Summary
A d.c. current-voltage (DCIV) technique is introduced in this chapter to investigate
hot-carrier-induced degradation in MOS transistors. This technique uses the base
current IB in a gate-controlled parasitic BJT of a MOST to monitor the stress-generated
oxide and interface traps. It has several unique features: (1) it is a pure d.c.
measurement which greatly simplifies instrumentation setup; (2) it is directly applicable
to submicron area devices in conventional production CMOS-based transistor structure
and can monitor the transistors degradation kinetics under hot-carrier stress; (3) the
magnitude of IB is directly proportional to the interface trap density NIT or QIX/q and
thus not affected by inhomgenity of NIT; (4) the surface recombination rate Rss and
minority injection level can be independently controlled by the applied gate voltage and
forward bias voltages, respectively; and (5) it has high sensitivity which gives a
minimum measurable NIT of < 109 traps/cm2 (Figure 3.8).
The DCIV technique has been recently packaged into a turn-key diskette
(SAMDCIV [71]) for the HP4156 (D.C. Parameter Analyzer), a typical characterization
equipment used among semiconductor companies. This diskette, together with a user-
guide and tutorial documentation, has been distributed to members of the
Semiconductor Research Corporation as part of a technology transfer effort.


108
known to occur near the drain junction of the nMOST. Typical degradation
characteristics of the quarter-micron nMOST under CHE stress are illustrated in Figures
8.5-8.7. Reduction of ISsat current, increases in VGT and IB.peak current are routinely
observed. In addition, the ISsat and VGT degradations are found to be linearly correlated
with the increase in the IB_peak current as shown in Figure 8.8. This indicates that
interface trap generation is a dominant degradation mechanism in these quarter-micron
low-voltage nMOSTs.
The measured current-accelerated and no-current-accelerated TTFs of the ISsat and
AVGt versus drain-source stress voltages VDS is shown in Figure 8.9(a). The failure
criteria used to determine the TTFs are defined as a 5% reduction in IDsat and a 10 mV
increase in AGT. A 120x stress-time reduction is observed for the nMOST stressed
under CACHE condition. In Figure 8.9(b), the FTF data are fitted to Sahs interface
trap theory using Eqn (7.12) where the bond-breaking energy qVBB in Eqn (7.12) is
replaced by an effective bond-breaking energy qVBB. to take into account of Auger-hot-
hole injection from the drain edge into the drain space-charge-region to generate
interface trap. As will be shown in Chapter 9, hot hole generated by Auger-
recombination of channel hot electron and thermal hole in the n+drain quasi-neutral
region requires the lowest electron kinetic energy. Hence VBB. is related to VBB by
^BB' ~ ^BB ^Si-Gap/^ (8.3)
where VR= VD VDsat and ESi.gap is the silicon energy gap. The interface trap
generation efficiency due to hot hole from Auger recombination pathway is then given
by


74
In Eqn.(5.9), the first subtraction used is to eliminate the contribution from the drain
surface recombination IB.2S component. The second subtraction used is to cancel out
the impact-generated hole current component. The final subtraction gives the stress
generated base current, AIB, as shown in Figure 5.4(b). It is observed that there is a
continuous rise of the IB (<~77pA) from the source towards the drain (~117pA) which
is expected from the channel hot electron stress. This AIB-VCB data is then least-square-
fitted to obtain the slope of 3(AIB)/3VCB which is shown in Figure 5.4(b). Evidently,
there is a change in slope in IB as VCB varies which indicates a non-uniform distribution
of stress-generated interface traps. The results qualitatively shows that the stress
generated interface trap starts to increase at about 0.6|im from the n+ drain junction
boundary( 9(AIB)/3VCB ~ 6pA/V), reaching about four times (~24pA/V) at 0.2pm from
the n+ drain junction boundary. This is consistent with the spatial variation of the
electric field where electric field strength increases after the depletion point, y =
y(VD.Sat)> and reaches a maximum at the drain metallurgical junction.
In summary, a simple and reliable extension of the DCIV technique is
demonstrated to profile the stress-generated interface trap densities along the channel of
MOS transistors.


78
negligible charging of the oxide traps over the drain-junction space-charge region
because the CHEs do not have sufficient kinetic energy, KE = q(YDS-VDS_sat) =
q[YDs-(VG-VGT)] = 3.3-1.6 = 1.7 eV, to surmount the 3.13 eV Si02/Si barrier. The
CHEs also do not have sufficient kinetic to create secondary hot holes, via interband-
impact electron-hole pair generation, to surmount the 4.25 eV Si02/Si hole barrier,
which would require a minimum energy of q(VDS-VDS.sat) = <|>B_hole Egap.Sj = (4.25 -
1.12) eV = 3.13 eV via the Auger generation pathway [117, 23]. The forward and
reverse IDsat and Id-Vgs characteristics of the nMOST is also shown in Figures 6.2 and
6.3, respectively. The asymmetrical difference between the post-stressed forward and
reverse characteristics clearly indicate the localized CHE-induced degradation near the
drain junction of the nMOST.
Figure 6.4 shows the linear dependence of AIs_sat and AVGT on AIB_peak for the two
sets of stress voltage conditions. This linear dependence is anticipated from the simple
textbook theory of ID_sat and VGT. Carrier mobility degradation in the channel should be
insignificant since the stress-generated traps are localized at the drain-junction. From
the simple parabolic MOST equation [99],
ID_sat = (W/2L)C0Hn(VGS VGT) 2, (6.1)
we get
^^D-sat^ ^D-satO = (2AVgt)/ (VgsVgt0) (6.2)
where the subscript 0 denotes the pre-stress value. Eqn.(6.2) assumes that (VG-VGT)
AVgt/2. Since AVGT = q(ANOT + ANIT)/C0X and IB NIT [59, 99-100], we can
immediately obtain AID.sat = AVGT <*= AIB.


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
CURRENT-ACCELERATED
CHANNEL HOT-CARRIER STRESS IN
SILICON METAL-OXIDE-SEMICONDUCTOR TRANSISTORS
by
Kim-Kwong Michael Han
December 1997
Chairman: Prof. Chih-Tang Sah
Major Department: Electrical and Computer Engineering
Transistor performance degrades during operation due to charging and generation of
oxide and interface traps by hot carriers. A novel direct-current current-voltage (DCIV)
technique is employed in this thesis to investigate the electrical degradation kinetics of
metal-oxide-semiconductor transistors (MOSTs) under hot-carrier stress. This
technique measures the interfacial recombination current IB (base or body current) as a
function of the gate voltage to monitor the generation rate of interface traps and
charging rate of oxide traps. Two DCIV applications are demonstrated: (1) separation
of stress-generated oxide and interface traps in submicron MOSTs, and (2) profiling the
spatial distribution of of the stress-generated interface traps. It is also shown that stress
generated surface recombination current peak is proportional to the traditionally
vi


CHAPTER 4
SEPARATION OF INTERFACE AND OXIDE TRAPS USING
THE DCIV METHOD
4.1 Introduction
Electrically active oxide (QOT) and interface traps (QIT) are known to affect the
electrical characteristics of metal-oxide-semiconductor capacitor (MOSC) and metal-
oxide-semiconductor transistors (MOST). Areally uniform and charged oxide traps in
MOSC and MOST will cause the Cg-VG and Iqs'^ds characteristics to shift in parallel
along the gate-voltage axis since the density of the oxide trap has no functional
dependence on the gate voltage, or the surface band bending condition. However, the
situation is more complex for interface traps because the density of charged interface
traps depends on the electron and hole concentrations at the Si02/Si interface which are
controlled by the applied gate voltage. Hence, interface traps will distort the Cg-VG and
ID-VG characteristics.
The traditional measurement techniques high-frequency Cg-VG (input capacitance)
and d.c. ID-VG (transfer characteristics) have been routinely used to study the generation
and charging of oxide and interface traps during hot carrier stress. For the two-terminal
high frequency Cg-VG method, the stress-created oxide trap density is monitored by the
change in flat-band voltage while the interface trap density is calculated from
experimental data using the Terman method. This method requires a large area MOSC
48


133
(VD VDsat)/(1 V)
Figure 9.8 Comparison of positive charge builup as a function of the electron kinetic
energy, (VD VDsat), at a stress fluence of Ninj= 20x1023 cm-2, from the
data given in Figures 9.3 and 9.6.


CH-electron /(1 A)
99
Figure 8.2 Effect of forward applied voltage to the bottom-emitter junction, VBE, on
the channel hot electron current versus drain voltage characteristics,
^CH-electron'^DS at VGS= 0 ^ V and VSB= 0


132
(VD VDsat)/(1 V)
Figure 9.7 Plot of positive charge builup as a function of the electron kinetic energy,
(VD VDsat), at a stress fluence of Ninj= 20x1 023 cm-2, from the data given
in Figure 9.6.


126
9.4.2 Experimental +0OT Build-up Threshold
The positive oxide charge build-up can be monitored by the gate-voltage shift of
the growing IB peak, VGB.peak, shown in Figure 9.1, since +AQOT c Cox.V GB.peak. This
negative gate-voltage shift of the growing IB peak would be proportional to the applied
drain voltage, VDS, since there are more hot holes with sufficient kinetic energy to
surmount the 4.25 eV Si02/Si hole barrier. Thus, a threshold is anticipated if VGB is
plotted against VDS provided the stress-fluence and other bias voltages are kept
constant. Figure 9.3 plots the variation of as a function of hot-electron fluence
for the L=100 |imxl00 (im transistors stressed with five different VDS voltages at the
common VGS= 0.9 V. The hot-electron fluence is calculated from the nearly-constant
stressed channel current IcH-stress ushig the following relationship.
NiNJ = (l/q)*I(IcH-stress^Drain)*dt (9.10)
- (l/q)(IGH-stress*tstress)^Drain
where q is the electronic charge (1.6xl019 Coulombs), ADrain is the drain contact area
and tstress is the incremental stress time. All five VGB_peak curves are observed to
saturate after Ninj= IOxIO23 cm-2. Ning reported a large capture cross section area
(~10~13cm2) for hole oxide traps in Si02 and observed an almost 99% hole trapping
efficiency [125]. This could possibly account for the VGBs saturation shown in Figure
9.3.
In Figure 9.3, the data corresponding to each VDS stress voltage was least-square-
fitted to two exponential functions to extract the VGB peak value at a constant hot-
electron fluence of Ninj= 20x10 23cm'2. The fitted results are shown as solid lines in


'Hito
92
Figure 7.2 Comparison of experimental and theoretical bond-breaking efficiency by
hot electrons and holes in Si npn BJT, nMOST and pMOST versus applied
junction voltage. 10-year TTF is the TTF at the applied junction voltage
indicated by the vertical arrow extrapolated by the theory from the TTF
data at the higher VR for 100% or 20% AIB. Adapted from Sah [42, p. 126]


118
captured by the oxide hole traps which becomes positively charged [125]. In addition,
charge pumping measurements also suggest that localized hole trapping into the gate
oxide can cause a negative voltage shift of charge-pumping characteristics [55, 122-
123]. A survey of the biasing conditions reported in [55,57-58, 113, 119-123] showed
high drain voltage (9.5 V employed to inject channel hot-holes into the Si02 oxide. However, the physical
mechanisms of how the hot hole are injected into Si02 during stress were not addressed
in the literature. Intuitively, if the energy of the hot hole is less than 4.25 eV, device
degradation caused by the injected channel hot-holes will not be a dominant factor
because the holes do not have sufficient kinetic energy to surmount the 4.25 eV Si02/Si
energy barrier [15, 50, 58]. Based on this reasoning, positive charge build up in Si02
will cease to occur. Hence, it is of fundamental interest and practical importance to
delineate and investigate the channel hot-hole injection mechanisms.
Recently, Lu and Sah showed two physical hot-hole back-injection pathways in
their thin-gate-oxide tunneling experiments [23], They considered the threshold energy
of two impact and two Auger pathways to account for the positive oxide-charge
buildup in thin Si02 [126]. We anticipate that their results can also be extended to
channel hot-hole injection.
In this chapter, four interband hot hole generation processes based on the
fundamental impact and Auger recombination pathways [126] are reviewed and
discussed. It will be shown, based on energy conservation, that there are three distinct
hot-electron (primary) threshold kinetic energies for generating a 4.25 eV hot hole


42
The surface electric field Es can be obtained as
E(x=0) = E(vs)
(3.31)
where Us, UF and UBE are the normalized potential with respect to kBT/q (Us =
q\(/s/kBT, UF = qVp/kBT, and UBE = qVBE/kBT).
The relationship between the applied gate voltage VGB and the surface band
bending potential \j/s is obtained by using the MOSC voltage equation.
(3.32)
VGB VFB Vs ^OX £SiEs/<-OX
where VFB is the flatband voltage, Vox is the oxide voltage, and Cox is the oxide
capacitance per unit area. Thus, for a given VGB, Eq.(3.32) can be solved iteratively for
\|/s. Three exact asymptotic solutions for Eq.(3.32) can be derived by weighting the
relative magnitude of the carrier concentrations at the Si02/Si interface under surface
accumulation, depletion, and inversion conditions [70].
Under surface accumulation condition where Us <0,
(VGB Vfb V's)
Vs = -(kBT/q)loge
+ A
A
(3.33)
(4kBT/q)VAA
Under surface depletion condition where 0< Us < 2UF UBE,
(^gb ^fb (kBT/q)AD
Vs ~ ^GB VpB+2VAA 1 1 +
(3.34)
Under surface inversion condition where Us > 2UF UBE,


6
stress. A hydrogen-diffusion model will be proposed to account for the reduction of
interface traps in the channel.
Chapter 11 gives the summaries and concludes this thesis.


143
The negative peak, starting at VGB= -0.5 V in Figures 10.1(b) and 10.1(d),
originates from the density-reduction of the residual fabrication interface traps located
in the channel during CHH stress. Its mid-channel location is deduced from its n=l
ideality factor, present at the higher forward bias (VSB= 0.6 V) that causes
recombination in the quasi-neutral region (signified by n=l) to dominate, and the near
symmetry between SE and DE. The slight asymmetry or the higher reduction rate in the
DE configuration, Figure 10.1(b), indicates the expected nearer p+drain location of the
mid-channel interface traps. The positive gate voltage shift of the decreasing peak
indicates that negative oxide charge build-up also occurs in the channel region.
10.3. Kinetic Model
The foregoing results on the generation of the interface traps and negative
charging of the oxide traps over the drain junction space-charge region are anticipated
from the similar results observed in nMOST during CHE stress. The same hydrogen
bond-breaking model would also apply. For the present pMOST case, interface traps
are generated when the hydrogen in SiH and SiOH at the Si02/Si interface over the
drain-junction space-charge region is released by the energetic secondary electrons
generated by the primary channel hot holes. Figures 10.3(a) and 10.3(b) show the
kinetics of the growth of the interface traps over in the drain-junction space-charge
region (n=2) which is initially limited by hydrogen-bond breaking with the linear stress
time dependence and then limited by hydrogen diffusion with square-root stress time
dependence. It saturates when all the hydrogenated silicon and oxygen dangling bonds
are dehydrogenated.


CHAPTER 2
STRESS-AND-MEASURE (SAM) METHODOLOGY
2.1 Introduction
In a Stress-And-Measure (SAM) experiment, the degradation characteristics of a
MOS transistor (MOST) subject to either substrate hot-carrier (SHC) or channel hot-
carrier (CHC) stress are monitored by the following electrical measurements: (1) drain-
current vs gate voltage (ID-VG), (2) saturation drain current vs gate voltage (losarlo)
(3) Direct-Current Current-Voltage (DCIV) or IB-VGB, and (4) forward-biased gated
emitter-base p/n junction (IB-VBE). These stress and measure steps are automated via
IEEE-488 bus in a SAM station. In this chapter, the SAM station setup and the
measurement and stress bias configurations are described. An overview of substrate-
and channel-hot electron stress is also presented.
2.2 SAM Setup
Figure 2.1 shows a block diagram of the SAM station setup. It consists of three
parts: (1) a Device-Under-Test (DUT) box, (2) an instrumentation rack of IEEE-488 bus
controlled power supplies, voltage-current pulse generators, digital current- and volt
meters, and (3) a Digital Equipment Cooporation MicroVax-II computer. The MOS
transistor (DUT), fabricated on a 6" or 8" wafer, is loaded into a light-shielded wafer
probe station where the Drain (D) Gate (G), Source (S), Body (X or B) and Substrate
(Sub or E for Emitter) of the DUT are connected via the probes and shielded low
7


12
HP3488
K230
J
Figure 2.4 SAM schematic setup for saturation IDsat measurement.


rTFOP/(1s)
105
IVDSI/(1V)
Figure 8.4 FTF (Fluence-to-Failure) on right y-axis and TTF0P (operation-TTF, i.e.
without current-acceleration) on left y-axis versus the magnitudes of VDS.
Solid curves are the LSF of the nMOST and pMOST data to the theory
giving bond-breaking kinetic energies of qVBB= 3.087 eV and 3.040 eV,
respectively.
FTF/(q/cm2)


154
[33] E. Takeda and N. Suzuki, "An Empirical Model for Device Degradation Due
to Hot-Carrier Injection," IEEE Electron Dev. Lett., Vol. EDL-4, No.4,
pp. 111-113, Apr. 1983.
[34] K.L. Chen, S.A. Sailer, I.A. Groves, and D. Scott, "Reliability Effects on MOS
Transistors Due to Hot-Carrier Injection," IEEE Trans. Electron Devices,
Vol.32, No.2, pp.386-393, Feb. 1985.
[35] C. Hu, S.C. Tam, F-C. Hsu, P.K. Ko, T-Y. Chan and K.W. Terrill, "Hot-
Electron Induced MOSFET Degradation Model, Monitor, and
Improvement," IEEE Trans. Electron Devices, Vol.32, No.2, pp.375-385, Feb.
1985.
[36] J.E. Chung, M.-C. Jeng, J.E. Moon, P.K. Po, and C. Hu, "Low-Voltage Hot-
Electron Currents and Degradation in Deep-Submirometer MOSFETs," IEEE
Trans. Electron Devices, Vol.37, No.7, pp.1651-1657, Jul. 1990.
[37] S.W. Sun, M. Orlowski and K.Y. Fu, "Parameter Correlation and Modeling of
the Power-Law Relationship in MOSFET Hot-Carrier Degradation," IEEE
Electron Dev. Lett., Vol.l 1, No.7, pp.297, 1990.
[38] Chih-Tang Sah, Fundamentals of Solid-State Electronics Study Guide, p.396-
397, World Scientific Publishing Co. Dec 1993. New Jersey and Singapore.
[39] A. Neugroschel, C.-T. Sah, and M.S. Carroll, "Current-acceleration for rapid
time-to-failure determination of bipolar junction transistors under emitter-base
reverse-bias stress," IEEE Trans. Electron Devices, Vol.42, No.7, pp.l 380-
1383, Jul. 1995.
[40] A. Neugroschel, C.-T. Sah, and M.S. Carroll, "Accelerated reverse emitter-
base bias stress methodologies and time-to-failure application," IEEE Electron
Device Lett., Vol.17, No.3, pp.l 12-114, March 1996.
[41] Arnost Neugroschel, Chih-Tang Sah, K. Michael Han, Michael S. Carroll,
Toshikazu Nishida, Jack T. Kavalieros and Yi Lu, "Direct-Current
Measurements of Oxide and Interface Traps on Oxidized Silicon," IEEE
Trans. Electron Devices, vol.42, no.9, Sept. 1995.
[42] Chih-Tang Sah, Fundamentals of Solid-State Electronics Solution Manual,
pp.l22-128, World Scientific Publishing Co. 1996. New Jersey and Singapore.
[43] Toshikazu Nishida and Scott E. Thompson, "Oxide field and thickness
dependence of trap generation in 9-30 nm dry and dry/wet/dry oxides, L
Appl. Phys., vol.69, pp.3986-3994, Apr. 1991.


56
Figure 4.4 Effect of areal-uniform SHEi stress on the n/p/n BJT and nMOST
characteristics as a function of VGB. All stressed at 5xl018 electron cm-2
fluence except bottom-E IB in (a). Measured at VEB= -0.3 V, VCB= 0 V
and 297K. (a) Base current and (b) collector current in the top-emitter and
bottom-emitter configurations, (c) Subthreshold Id-Vgb of the nMOST.


69
VGB/(1V)
Figure 5.2 Effect of channel hot electron stress on the Drain-Emitter and Source-
Emitter DCIV characteristics. All measured at VBE= 0.4 V, VCB= 0.0 V
and 297K. CHE stress conditions are VDS= 8.0 V and VGS= 4.0 V for 120
minutes. The solid line refer to the post-stressed measurements while the
dotted line is for the unstressed measurement. The unstressed IB-VGB
measurement is similar for both drain and source junctions and hence only
one curve is shown.


39
Â¥s/(1V)
Figure 3.6 Relation of Rss and \jrs with VBE as parameters. Assumed mid-energy-
level interface trap in Si energy gap and cns=cps. NAA= 1016 cm-3. T=300
K.


122
energies derived analytically based on energy and momentum conservation are nearly
equal to their theoretical values computed from energy conservation alone which are
due to the electron-intervalley and umklapp transitions in silicon [109],
9.3 Experiments
Two groups of n-channel MOS transistor structures with two different drawn
channel lengths (L=100 |im and 20 pm) are used. The MOS transistor with L= 100 |im
is fabricated from a 1.0 (im BiCMOS technology with junction well isolation and has an
130 A thick gate oxide and a drawn channel width of 100 |im. The MOS transistor with
L=20 |im is fabricated from a 0.5 |im CMOS technology and has 80A thick gate oxide
and 20 |im channel width.
For a given SAM run on a test transistor, the transistor was stressed for a large
number of stress intervals at a constant high VDS.stress and a common low VGS_stress.
The VDS_stress was different for each transistor in order to vary the kinetic energy of the
primary channel electron since KEe|ectron = q(VDS VDsat) is the maximum kinetic
energy gained by the channel electron accelerating from the low-field edge of the drain
junction space-charge-layer to the edge of the drain n+/p metallurgical junction. After
each stress interval, the transistor characteristics were measured using the DCIV method
in the drain-emitter configuration (DE-DCIV) to monitor the stress-generated NIT and
+Qot-


-12
77
Figure 6.
Drain-Emitter DE-DCIV characteristics of a nMOST stressed by channel
hot electrons.


152
[10] Y. Taur, D.A. Buchanan, W. Chen, D.J. Frank, K.E. Ismail, S.-H. Lo, G.A.
Sai-Halasz, R.G. Viswanathan, H.-J.C. Wann, S.J. Wind, and H-S Wong,
"CMOS Scaling into the Nanometer Regime," Proceeding of the IEEE.
Vol.85, No.4, pp.486-504, Apr. 1997.
[11] R.H. Dennard, F.H. Gaensslen, H.-N. Yu, V.L. Rideout, E. Bassous and A.R.
LeBlanc, "Design of Ion-Implanted MOSFETs with Very Small Physical
Dimensions," IEEE J. Solid-State Circuits, Vol.SC-9, No.5, pp.666, 1974.
[12] J.R. Brews, W. Fichtner, E.H. Nicollian, and S.M. Sze, "Generalized Guide for
MOSFET Miniaturization," IEEE Electron Dev. Lett., Vol.l, No.l, pp.2-4,
Jan. 1980.
[13] P.K. Chatterjee, W.R. Hunter, T.C. Holloway and Y.T. Lin, "The Impact of
Scaling Laws on the Choice of n-Channel or p-Channel for MOS VLSI,"
IEEE Electron Dev. Lett., Vol. 10, No. 10, pp.220-223, Oct. 1980.
[14] G. Baccarani, M.R. Wordeman, and R.H. Dennard, "Generalized Scaling
Theory and Its Application to a 1/4 Micrometer MOSFET Design," IEEE
Trans. Electron Devices, Vol.31, No.4, pp.452-462, Apr 1984.
[15] C.T. Sah, "Models and Experiments on Degradation of Oxidized Silicon,"
Solid-State Electronics, Vol. 33, No.2, pp.147, 1990.
[16] J.F. Verwey, "Nonavalanche Injection of Hot Carriers into Si02," J. Appl.
Phvs., Vol.44, No.6, pp.2681, 1973.
[17] T.H. Ning, "Hot-Electron Emission from Silicon into Silicon Dioxide," Solid-
State Electronics, Vol.21, pp.273, 1978.
[18] E.H. Nicollian and C.N. Berglund, "Avalanche Injection of Electrons into
Insulating Si02 Using MOS Structures," J. Appl. Phys., Vol.41, No.7,
pp.3052, 1970.
[19] M. Lenzinger and E.H. Snow, "Fowler-Nordheim Tunneling into Thermally
Grown Si02," J. Appl. Phys., Vol.40, No.l, pp. 278, 1969.
[20] S.E. Thompson and T. Nishida, "Tunneling and Thermal Emission of
Electrons from a Distribution of Shallow Traps in Si02," Appl. Phys. Lett.,
Vol.58, No.12, pp. 1262, 1991.
[21] D.J. DiMaria and J.W. Stasiak, "Trap Creation in Silicon Dioxide Produced by
Hot Electrons," J. Appl. Phys., Vol.65, No.6, pp.2342, 1989.


153
[22] Charles C.H. Hsu, T. Nishida, and Chih-Tang Sah, "Observation of Threshold
Oxide Electric Field for Trap Generation in Oxide Films on Silicon," J. Appl.
Physic, Vol.63, No.12, 15 June 1988.
[23] Yi Lu and Chih-Tang Sah, "Two Pathways of Positive Oxide-Charge Buildup
During Electron Tunneling into Silicon Dioxide Film," J. Appl. Phys., vol.76,
pp.4724-4727, Oct. 1994.
[24] Scott E. Thompson and Toshikazu Nishida, "Positive charge generation in
Si02 by electron-impact emission of trapped electrons," J. Appl. Phys., vol.72,
no. 10, pp.4683-4695, Nov. 1992.
[25] S.K. Lai and D.R Young, "Effects of Avalanche Injection of Electrons into
Silicon Dioxide Generation of Fast and Slow Interface States," J. Appl.
Phys., vol.52, no. 10, pp.6231-6240, Oct. 1981.
[26] D.A. Buchanan and D.J. DiMaria, "Interface and Bulk Trap Generation in
Metal-Oxide-Semiconductor Capacitors," J. Appl. Phys., vol.67, no. 12,
pp.7439-7452, Jun. 1990.
[27] M.V. Fischetti, "Generation of Positive Charge in Silicon Dioxide During
Avalanche and Tunnel Electron Injection," J. Appl. Phys., vol.57, no.8,
pp.2860-2879, Apr. 1985.
[28] Chih-Tang Sah, "Fundamental of solid-state electronics," World Scientific
Publishing Co,. Singapore 1991. See Section 684 "High Electric Field and
High Voltage Effects".
[29] E. Takeda, H. Kume, T. Toyane and S. Asai, "Submirometer MOSFET
Structure for Minimizing Hot-Carrier Generation," IEEE Trans. Electron
Devices, Vol. ED-29, No.4, pp.611, 1982.
[30] S. Ogura, P.J. Tsang, W.W. Walker, D.L. Critchlow and J.F. Shepard, "Design
and Characteristics of the Lightly Doped Drain-Source (LDD) Insulated Gate
Field-Effect Transistor," IEEE Trans. Electron Devices, Vol.ED-27, No.8,
pp. 1359, 1980.
[31] J.J.Sanchez, K.K. Hsueh, T.A DeMassa, "Drain-Engineered Hot-Electron-
Resistant Device Structures: A Review," IEEE Trans. Electron Devices,
Vol.36, No.6, pp.l 125-1132, Jun 1989.
[32] F.C. Hsu, "Hot-Carrier-Resistant Structures," Chapter 4, VLSI Electronics
Microstructure Science, Vol. 18, Advanced MOS Device Physics, Academic
Press, 1989.


AIb-de/(10 12 A)
136
VGB/(1V)
Figure 9.9 Drain-Emitter DCIV of two 0.35 |lm nMOSTs, Transistor-A stressed at
VDS= 5.0 V for tstress= 11.7 min and Transistor-B stressed at VDS= 3.3 V
for tstress= 494.8 min. VGS.stress= 0.5 V. The dotted lines shows the
unstressed DE-DCIV characteristics of Transistor-A and Transistor-B.
Ib-de/(10 12 A)


160
[97] V.G.K. Reddi, "Influence of Surface Conditions on Silicon Planar Transistor
Current Gain," Solid-State Electronics, vol.10, pp.305-334, 1967.
[98] Chih-Tang Sah, "Fundamental of solid-state electronics," World Scientific
Publishing Co,. Singapore 1991. See Eqn.(531.1) on p.424.
[99] Chih-Tang Sah, "Fundamental of solid-state electronics," World Scientific
Publishing Co,. Singapore 1991. See (643.6) for ID_sat and (682.20) and
(683.9) for SocDit.
[100] J.F. Verwey, "The Introduction of Charge in Si02 and The Increase of
Interface States During Breakdown of Emitter-Base Junction of Gated
Transistors," Appl. Phys. Lett., vol.15, no.8 pp.270-272, Oct. 1969.
[101] Chih-Tang Sah, "Interface Traps on Si Surface," EMIS Datareview
RN=16170, section 17.1, p.499, 1987.
[102] F.-C. Hsu and S. Tam, "Relationship between MOSFET Degradation and Hot-
Electron-Induced Interface-State Generation," IEEE Electron Devices Letters,
vol.5, no.2, pp.50-52, Feb. 1984.
[103] K.R. Hofmann, C. Werner, W. Weber, and G. Dorda, "Hot-Electron and Hole-
Emission Effects in Short n-Channel MOSFETs," IEEE Trans. Electron
Devices, Vol.32, No.3, pp.691-699, Mar 1985.
[104] E. Takeda, A. Shimizu, and T. Hagiwara, "Role of Hot-Hole Injection in Hot-
Carrier Effects and the Small Degraded Channel Region in MOSFETs,"
IEEE Electron Devices Letters, vol.4, no.9, pp.329-331, Sep. 1983.
[105] H. Gesch, J-P Leburton, and G. Dorda, "Generation of Interface States by Hot
Hole Injection in MOSFETs," IEEE Trans. Electron Devices, Vol.29, No.5,
pp.913-918, May 1982.
[106] R. Fair and R.C. Sun, "Threshold-Voltage Instability in MOSFETs Due to
Channel-Hot Hole Emission," IEEE Trans. Electron Devices, Vol.28, No.l,
pp.83-94, Jan. 1981.
[107] S.K. Lai, "Two-Carrier Nature of Interface-State Generation in Hole Trapping
and Radiation Damage," Appl. Phys. Lett., Vol.39, No.l, pp.58-60, July 1981.
[108] S.K. Lai, "Interface Trap Generation in Silicon Dioxide When Electrons Are
Captured by Trapped Holes," J. Appl. Phys., Vol. 54, No.5, pp.2540-2546,
May 1983.


28
Source Gate Drain Base
1
Si20 Q,t oQoT 1
A A A__ A A
. l
n: n
+ p+base
contact

i ;
p-base well
n-epitaxial layer
n+ substrate
I
Substrate
Figure 3.1 Cross-sectional view of a n-channel MOST showing the lateral and vertical
parasitic n/p/n BJT.


CHAPTER 10
REDUCTION OF INTERFACE TRAPS IN P-CHANNEL MOS TRANSISTOR
DURING CHANNEL-HOT-HOLE STRESS
10.1 Introduction
The dominant degradation mechanism in p-channel MOS transistors (pMOSTs)
under channel hot hole (CHH) stress at low VGS and high VDS has been attributed to
build-up of negative oxide charge (-Qqt) due to electron trapping in the gate oxide near
the drain junction space-charge region [51, 127]. Reliability modeling of pMOST has
assumed channel-shortening caused by this localized negative oxide charge [128-130],
Few studies have investigated the mechanisms of interface trap (NIT) generation and its
effects on the performance and degradation of pMOSTs [55, 130-131]. In this chapter,
we report not only the generation of interface traps but also reduction of interface traps
for the first time in pMOSTs during CHH stress. Based on a variety of geometrical and
bias dependences on the experimental results, a hydrogen reaction-transport model is
constructed.
10.2. Experimental Results
The test transistors used in the stress-and-measure (SAM) experiments were
pMOSTs in n-well on p-substrate with 110A gate oxide and a drawn aspect ratio of
W/L= 20 p.m/0.7 Jim. They were stressed at room temperature under the conventional
CHH condition: VDS= -5.25 V, VGS= -1.0 V, Vsx= 0 V. Generation and reduction of
interface traps and charging of the oxide traps are monitored by the DCIV method. Two
138


115
VDS/(1V)
Figure 8.9 (a) Time-to-Failure versus drain stress voltage of 17.5(imx0.35[j.m nMOST
for no-current-acceleration stress (open triangles and circles) and current-
accelerated stress (solid triangles and circles). All transistors are stressed at
Vgs=0.5V. IDsat is measured at VD=VG= 1.0 V and VGT is extracted at ID=
0.5 mA.
(b) FTF (Fluence-to-Failure) on right y-axis and TTFOP (operation-TTF,
i.e. without current-acceleration) on left y-axis versus the magnitudes of
VDS. Solid curve is the LSF of the nMOST IDsat data giving bond-breaking
kinetic energy of qVBB= 3.054 eV.
FTF/(q/cm2)


(V 9-01- )/a I
130
VDS/(1 V)
Figure 9.5 ID-VD characteristics of the 100(imx 1 OOfim transistor measured at VGS=
0.9V. The drain conductance, gD, is determined from gD= AID/AVD. The
zero VDS-intercept by the linear-LSF line of gD gives VDsat= 0.3314 V.
90/(10


137
stress will cease. It is also inferred that interface trap would be a primary degradation
factor affecting the reliability of low-voltage deep submicron MOST where Q0t
buildup is no longer important.


119
(secondary). Experiment will show that, during channel hot electron stress, a 4.25 eV
interband impact-generated secondary hot hole is created when a primary impacting
electron possesses kinetic energy of more than 5.37 eV.
9.2 Four Fundamental Interband Hot Hole Generation Pathways
Figure 9.1 shows four fundamental interband generation pathways to create a 4.25
eV hot hole with the hot electron as the initiating carrier. For each pathway shown in
Figure 9.1, the energy of the hot electron and holes is drawn to-scale. This systematic
classification of generating 4.25 eV hot hole by the primary hot electron was delineated
by Sah based on the fundamental impact and Auger transitions [126].
Pathway (1) shows a 4.25 eV hot hole is generated by Auger recombination of a
hot electron with a thermal hole. The neighboring hole is excited to higher energy after
absorbing the recombination energy. From energy conservation, the threshold electron
kinetic energy needed to generate a 4.25 eV hot hole is
^electron-1 = t'X-h ~ EG (9-1)
= (4.12- 1.12) eV
= 3.13 eV.
where <[)x_h is the Si02/Si hole barrier height and EG is the silicon energy gap.
Pathway (2) shows another hot hole generation process by Auger recombination of
a hot electron with a thermal hole with a deep valence electron excited into the
conduction band edge to carry away the recombination energy. The threshold kinetic
energy of the initiating KEe|ectron.2 required to generate a 4.25 eV hot hole is given by


K230
K230
J
Figure 2.3 SAM schematic setup for linear ID-VG measurement.


94
quality of the Si02 oxide passivating the Si02/Si interface. For MOST, this oxide is
thermally grown on the Si substrate which is inherently better than the chemically-
deposited oxide used to cover the emitter-base interface of the npnBJT. Difference in
the surface impurity concentration profile associated with the fabrication of the npnBJT
and the MOST may also affect the bond-breaking energy.


155
[44] T.H. Ning and H.N. Yu, "Optically Induced Injection of Hot Electrons into
Si02," J. Appl. Phvs, Vol.45, No. 12, pp.5373, 1974.
[45] W. Shockley, "Problems Related to p-n Junctions in Silicon," Solid-State
Electronics, Vol.2, No.l, pp.35, 1961.
[46] T.H. Ning, C.M. Osburn and H.N. Yu, "Emission Probability of Hot Electrons
from Silicon into Silicon Dioxide," J. Appl. Phys., Vol.48, pp.286, 1977.
[47] C. Hu, "Lucky-Electron Mode of Channel Hot Electron Emission," IEDM,
pp.22, 1979.
[48] S. Tam, P.K. Ko and C. Hu, "Lucky-Electron Model of Channel Hot-Electron
Injection in MOSFETs," IEEE Trans. Electron Devices, Vol. ED-31, No.9,
pp.1116, 1984.
[49] S.A. Abbas and R.C. Dockerty, "Hot-Carrier Instability in IGFETs," Appl.
Phys. Lett., Vol.27, No.3, pp.147-148, Aug 1975.
[50] P.E. Cottrel, R. Troutman, and T.H. Ning, "Hot-Electron Emission in N-
Channel IGFETs," IEEE Trans. Electron Devices, vol.26, no.4, pp.520-533,
Apr. 1987.
[51] K.K. Ng and G.W. Taylor, "Effects of Hot-Carrier Trapping in n- and p-
Channel MOSFETs," IEEE Trans. Electron Devices, Vol.30, No.8, pp.871-
876, Aug.v 1983.
[52] B. Doyle, M. Bourcerie, J.-C. Marchetaux and A. Boudou, "Interface State
Creation and Charge Trapping in the Medium-to-High Voltage Range (Vd/2
Vg Vd) During Hot-Carrier Stressing of n-MOS Transistors," IEEE Trans.
Electron Devices, Vol.37, No.3, pp.744-754, Mar. 1990.
[53] J.Y. Choi, P.K. Po, C. Hu and W. Scott,"Hot Carrier-Induced Degradation of
Metal-Oxide-Semiconductor Field Effect Transistors: Oxide Charge Versus
Interface Traps," J. Appl. Phys., Vol.65, No.l, pp.354-360, Jan. 1989.
[54] S.A. Abbas, "Substrate Current A Device and Process Monitor," IEDM, 17.7,
pp.404-407, Dec. 1974.
[55] Paul Heremans, Rudi Bellens, Guido Groeseneken and Herman E. Maes,
"Consistent Model for the Hot-Carrier Degradation in n-channel and p-
Channel MOSFETs," IEEE Trans. Electron Devices, Vol.35, No. 12, pp.2194-
2209, Dec. 1988.


127
Figure 9.3 Kinetics of positive oxide charge buildup versus stress fluence, Ninj, in 100
pm x 100 Jim nMOSTs stressed at VDS= 5.5V, 6.0V, 6.5V and 7.0V, and
the same VGS= 0.9V


(sO/dli
102
VDS/(1V)
Figure 8.3 (b) Time-to-Failure versus drain stress voltage of 20jimx20|im pMOST for
no-current-acceleration stress (four circles) and current-accelerated stress
(five solid dots at VEB= 0.85 V). All transistors are stressed at VGS= -1.0
V. The inset shows the DE-DCIV (IB-VGB) characteristics (measured at
Vdb= 0.4 V, VSB=VEB= 0.0 V) as a function of CHH stress time (0, 85.3,
738.7, 2319.7, 17227.6, and 1.8xl05 sec). Stressed at VDS= -5.25 V, VGS=
-1.0 V, and VSB=VBE=0 V.


104
n+/p junction (one asterisks at VBS= 0.7 V for VDS= 6.0 V giving about 16x
acceleration). At each applied VDS_stress, the applied VGS.stress in (2) and (3) was kept
the same as in (1). Reduction in SAM times to measure TTF using CACHE scales with
the increased CHE current by forward-biasing the substrate/body or source/body n+/p
junction. For example, at VDS.stress= 6.0 V and VBE= 0.63 V, TTFnoCACHE/TTFCHE =
19200s/1240s = 16 and ICH-electron(VBE= 0-63 V)/ICH.e,ectron(VBE= 0 V) = 189.8 flA/16.8
pA = 12.3. Without current-acceleration, the estimated SAM time at VDS.stress=4.5V is
= 8x107s = 950 days (unfilled triangle) from extrapolating the four no-CACHE TTF
data points (circles) to 4.5V. Current acceleration via VBE= 0.74 V gave a 50x
reduction of the stress test time to 1.6xl06 s = 18.5 days (filled triangle). Thus, this
current-acceleration experiment has demonstrated the practical application in reducing
the measurement time from 950 days or 3 years to 19 days or 3 weeks in order to
determine the 10-year operation TTF of this transistor technology.
The CACHH TTF data of the pMOST in Figure 8.3(b) were obtained by forward
biasing the substrate/body p/n junction to VBE= -0.85 V, which showed only a lOx
reduction. This is smaller than the acceleration observed in nMOST in Figure 8.3(a) at
a lower forward biases, VBE= 0.63 V (16x) and 0.74 V (50x) and it is attributed to the
larger series base resistance of the n-well and p-substrate which reduced the forward-
current.
The TTFqps in Figure 8.4 at the designed lower operation voltages with no
current acceleration were obtained from the accelerated and not-accelerated fluence-to-


54
3.12V. In the SHEi experiment, a reverse bias of 4.0V is chosen for VCB. Some of the
electrons that are injected into the Si02 can be captured by the neutral oxygen vacancy
centers [72], giving
VG + e -> V0- (4.3)
and results in a net built-up of negative Q0T. Because of the high kinetic energy (~4eV
from Vsb=Vdb=4.0V) of the accelerating electrons in the surface space-charge-layer,
they can also break the weak or strained interfacial bonds (Si-Si, Si-O, Si-H, SiO-H) via
direct impact to create new interface traps [15] since the bond energy of of these
interfacial bond is of the order of 3eV.
During SHEi, the gate oxide current is kept constant, within 5% of the initial
preset gate current, by digital feedback control of the emitter/base forward bias. The
oxide electric field, Eox, is given by (VG Vs)/xox. The gate electron fluence, FG or
Ninj, is calculated as follows,
Ninj = J* (JG/q) dt (4.4)
= (JG/q)*t stress (4.5)
where JG is the constant gate current density, q is the electronic charge (l.xlO'19 C),
and tstress is the total stress time in seconds.
One key feature of using the SHEi technique is that the three basic stress
parameters, the electron kinetic energies in the Si02 and at the Si02/Si interface, and the
injection current can be independently controlled [20, 22, 43], The electron kinetic
energy in the Si02 is controlled by the voltage applied between the gate terminal and the
n+source/n+drain terminals provided the gate voltage is sufficient to invert the silicon


29
During DCIV measurement, the emitter-base junction is forward-biased and the
collector-base junction is short-circuited. The gate-base voltage is stepped from
accumulation to inversion. The base or body current IB is measured using a
picoammeter at each gate voltage step. Thus, a DCIV measurement is a plot of the base
or body terminal current versus the gate-base or gate-body voltage at a constant
forward-biased emitter-base voltage.
3.3 Components of IB
During hot-carrier stress, the interface traps (open triangles in Figure 3.2) along the
Si02/Si interface can be generated over the drain/body junction space-charge-region and
over the n-channel region from hot carriers breaking the strained or weak hydrogen
bonds (Si:H and SiO:H). Thus, during DE-DCIV measurement, the measured IB_DE
contains two emitters (not shown in Figure 3.2) and six base recombination components
[59-60, 63-65], The two emitter components are the electron-hole (e-h) recombinations
at the recombination centers in the drains bulk-quasi-neutral region (Ig-ei-buik) ar>d at
the interface between poly-silicon and crystalline-silicon contacts (IB.ei-contact)>
respectively. The base components are due to e-h recombinations centers or interface
traps located in the six regions. In Figure 3.2, these six base recombination components
are distinguished by one alphabetic and one numeric subscript: c^channel region,
sssurface space-charge region, b=bulk space-charge region; 1 and 2 from IB =
exp(qVBE/nkBT) where n=l for Shockleys exact p/n junction diode law and n=2 for the
simplified Sah-Noyce-Shockley (SNS) space-charge-layer recombination law [59, 65].
These six components are: (1) Ib_de-c-1 over n-channel-interface, (2) Ib_de-s-2 over


139
bias configurations were employed in the DCIV measurements: Drain-Emitter and
Source-Emitter forward bias (DE-DCIV and SE-DCIV) to locate the interface and oxide
traps. The trap locations are further delineated by the ideality factor of the I-V
characteristics of the forward-biased junction [59].
Figure 10.1 shows typical DE-DCIV characteristics of a CHH-stressed 20/0.7
pMOST where the stress-generated incremental base current is computed from
experimental data using
(^stress 1> ^EBl) ~ ^B(Wess 1 ^EBl) ~ ^B^stress=0 ^EBl)- (10.1)
The incremental rather than total body (or base) current gives visibility of the small
CHH-stress-induced changes, such as the decreasing peak at VGB= -0.5 V in Figures
10.1(b) and 10.1(d) and the increasing-then-decreasing peak at VGB= +2.0 V in Figure
10.1(c), that reveal for the first time electrical-annealing of the interface traps during
CHH stress. These small peaks would be masked off by the pre-stress leakage current if
the total IB were plotted. In addition, the incremental base current also subtracts out the
base/body-current components from bulk and interface traps outside of the pMOST
channel which are not changed during the CHH stress and which may also not be
dependent on the gate voltage.
The DCIV data in Figure 10.1 are displayed at two forward biases (VEB= 0.4 V
and 0.6 V) for the two bias configurations (DE and SE) shown in Figures 10.1 (a)-(d).
These four DCIV plots help to locate the interface and oxide traps. At the lower
forward bias (0.4 V), AIB is dominated by electron-hole recombination at the interface
traps located over the surface space-charge-region of the forward-biased drain or source



PAGE 1

&855(17$&&(/(5$7(' &+$11(/ +27&$55,(5 675(66 ,1 6,/,&21 0(7$/2;,'(6(0,&21'8&725 75$16,67256 %< .,0.:21* 0,&+$(/ +$1 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$

PAGE 2

$&.12:/('*(0(176 ZLVK WR H[SUHVV P\ JUDWLWXGH WR 3URIHVVRU &KLK7DQJ 6DK IRU DOO KLV ZLVGRP JXLGDQFH SDWLHQFH DQG WLPH WKURXJKRXW WKH FRXUVH RI P\ JUDGXDWH UHVHDUFK DW WKH 8QLYHUVLW\ RI )ORULGD ZRXOG DOVR OLNH WR H[SUHVV P\ WKDQNV WR 3URIHVVRUV $UQRVW 1HXJURVFKHO 7RVKLND]X 1LVKLGD 3HWHU =RU\ DQG 8OULFK .XU]ZHJ IRU VHUYLQJ RQ P\ VXSHUYLVRU\ FRPPLWWHH ZLVK WR WKDQN P\ SDVW DQG SUHVHQW FROOHDJXHV 'U 6FRWW 7KRPSVRQ 'U
PAGE 3

7$%/( 2) &217(176 3DJH $&.12:/('*(0(176 LL $%675$&7 YL &+$37(56 ,1752'8&7,21 675(66$1'0($685( 6$0f 0(7+2'2/2*< ,QWURGXFWLRQ 6$0 6HWXS (OHFWULFDO 0HDVXUHPHQW 6XEVWUDWH DQG &KDQQHO +RW (OHFWURQ 6WUHVV ',5(&7&855(17 & 8 5 5 ( 17 9 2 / 7 $ ( '&,9f &+$5$&7(5,=$7,21 7(&+1,48( ,QWURGXFWLRQ 0HDVXUHPHQW &RQILJXUDWLRQV &RPSRQHQWV RI ,% 7KHRU\ RI 6XUIDFH 5HFRPELQDWLRQ 5DWH 5HODWLRQVKLS EHWZHHQ ?_V DQG 9*% XQGHU 1RQ(TXLOLEULXP &RQGLWLRQ 6XPPDU\ 6(3$5$7,21 2) ,17(5)$&( $1' 2;,'( 75$36 86,1* 7+( '&,9 7(&+1,48( ,QWURGXFWLRQ ([SHULPHQW 5HVXOWV DQG 'LVFXVVLRQV LLL

PAGE 4

352),/,1* 2) ,17(5)$&( 75$36 %< 7+( '&,9 0(7+2' ,QWURGXFWLRQ 7KHRU\ ([SHULPHQW DQG 5HVXOWV /,1($5 5('8&7,21 2) '5$,1 &855(17 :,7+ ,1&5($6,1* ,17(5)$&( 5(&20%,1$7,21 ,1 1026 75$16,67256 675(66(' %< &+$11(/ +27 (/(&75216 ,QWURGXFWLRQ ([SHULPHQWV 5HVXOWV DQG 'LVFXVVLRQ ,17(5)$&( 75$36 *(1(5$7,21 02'(/ ,QWURGXFWLRQ 7KHRU\ ([SHULPHQWV DQG 5HVXOWV 3+<6,&6%$6(' 7,0(72)$,/85( (;75$32/$7,21 $/*25,7+0 86,1* &855(17$&&(/(5$7(' &+$11(/ +27&$55,(5 675(66 ,QWURGXFWLRQ 0HWKRGRORJ\ RI &XUUHQW$FFHOHUDWHG &KDQQHO+RW&DUULHU 6WUHVV 5HVXOWV DQG 'LVFXVVLRQV +27 +2/( ,1-(&7,21 ,172 6L ,1 1&+$11(/ 026 75$16,6725 '85,1* &+$11(/+27(/(&7521 675(66 ,QWURGXFWLRQ )RXU )XQGDPHQWDO ,QWHUEDQG +RW +ROH *HQHUDWLRQ 3DWKZD\V ([SHULPHQWV 5HVXOWV DQG 'LVFXVVLRQV 5('8&7,21 2) ,17(5)$&( 75$36 ,1 3&+$11(/ 026 75$16,6725 '85,1* &+$11(/+27+2/( 675(66 ,QWURGXFWLRQ ([SHULPHQWDO 5HVXOWV .LQHWLF 0RGHO 6XPPDU\ ,9

PAGE 5

6800$5,(6 $1' &21&/86,216 5()(5(1&(6 %,2*5$3+,&$/ 6.(7&+ Y

PAGE 6

$EVWUDFW RI 'LVVHUWDWLRQ 3UHVHQWHG WR WKH *UDGXDWH 6FKRRO RI WKH 8QLYHUVLW\ RI )ORULGD LQ 3DUWLDO )XOILOOPHQW RI WKH 5HTXLUHPHQWV IRU WKH 'HJUHH RI 'RFWRU RI 3KLORVRSK\ &855(17$&&(/(5$7(' &+$11(/ +27&$55,(5 675(66 ,1 6,/,&21 0(7$/2;,'(6(0,&21'8&725 75$16,67256 E\ .LP.ZRQJ 0LFKDHO +DQ 'HFHPEHU &KDLUPDQ 3URI &KLK7DQJ 6DK 0DMRU 'HSDUWPHQW (OHFWULFDO DQG &RPSXWHU (QJLQHHULQJ 7UDQVLVWRU SHUIRUPDQFH GHJUDGHV GXULQJ RSHUDWLRQ GXH WR FKDUJLQJ DQG JHQHUDWLRQ RI R[LGH DQG LQWHUIDFH WUDSV E\ KRW FDUULHUV $ QRYHO GLUHFWFXUUHQW FXUUHQWYROWDJH '&,9f WHFKQLTXH LV HPSOR\HG LQ WKLV WKHVLV WR LQYHVWLJDWH WKH HOHFWULFDO GHJUDGDWLRQ NLQHWLFV RI PHWDOR[LGHVHPLFRQGXFWRU WUDQVLVWRUV 0267Vf XQGHU KRWFDUULHU VWUHVV 7KLV WHFKQLTXH PHDVXUHV WKH LQWHUIDFLDO UHFRPELQDWLRQ FXUUHQW ,% EDVH RU ERG\ FXUUHQWf DV D IXQFWLRQ RI WKH JDWH YROWDJH WR PRQLWRU WKH JHQHUDWLRQ UDWH RI LQWHUIDFH WUDSV DQG FKDUJLQJ UDWH RI R[LGH WUDSV 7ZR '&,9 DSSOLFDWLRQV DUH GHPRQVWUDWHG f VHSDUDWLRQ RI VWUHVVJHQHUDWHG R[LGH DQG LQWHUIDFH WUDSV LQ VXEPLFURQ 0267V DQG f SURILOLQJ WKH VSDWLDO GLVWULEXWLRQ RI RI WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDSV ,W LV DOVR VKRZQ WKDW VWUHVVn JHQHUDWHG VXUIDFH UHFRPELQDWLRQ FXUUHQW SHDN LV SURSRUWLRQDO WR WKH WUDGLWLRQDOO\ YL

PAGE 7

PRQLWRUHG WUDQVLVWRU GHJUDGDWLRQ SDUDPHWHUV GUDLQVDWXUDWLRQ FXUUHQW JDWH WKUHVKROG YROWDJH DQG VXEWKUHVKROG YROWDJH VZLQJ $,'VDW $9*7 DQG $6f $ SK\VLFVEDVHG 7LPHWR)DLOXUH H[WUDFWLRQ DOJRULWKP LV GHYHORSHG XVLQJ WKH QHZ FXUUHQWDFFHOHUDWHG FKDQQHOKRWFDUULHU VWUHVV &$&+&f PHWKRG &XUUHQWDFFHOHUDWLRQ RI WKH WUDQVLVWRUVf GHJUDGDWLRQ UDWH LV REWDLQHG E\ IRUZDUGELDVLQJ D SQ MXQFWLRQ WR LQFUHDVH WKH FKDQQHO KRW FDUULHU FXUUHQW 7KH DSSOLHG VWUHVV YROWDJHV FDQ EH LQGHSHQGHQWO\ VHW WR FRQWURO WKH NLQHWLF HQHUJ\ RI WKH FKDQQHO KRW FDUULHUV 6WUHVVWLPH UHGXFWLRQV E\ DERXW RUGHUV RI PDJQLWXGH DUH GHPRQVWUDWHG &RUUHODWLRQ ZLWK 6DKfV ERQGEUHDNLQJ LQWHUIDFH JHQHUDWLRQ PRGHO VKRZHG WKDW LQWHUIDFLDO VLOLFRQK\GURJHQ ERQGV DUH EURNHQ E\ KRW KROHV LQ QFKDQQHO 0267 DQG E\ KRW HOHFWURQV LQ SFKDQQHO 0267 3RVLWLYH FKDUJLQJ RI R[LGH WUDS E\ EDUULHUVXUPRXQWLQJ KRW KROH GXULQJ FKDQQHO KRW HOHFWURQ VWUHVV LV LQYHVWLJDWHG XVLQJ WKH '&,9 WHFKQLTXH 7ZR SDWKZD\V DUH H[SHULPHQWDOO\ REVHUYHG f KRWHOHFWURQKRWKROH LQWHUEDQG LPSDFW SDWKZD\V LQ KLJK YROWDJH ORQJFKDQQHO Q0267 ZLWK HOHFWURQ NLQHWLF HQHUJ\ WKUHVKROG RI H9 DQG f KRWHOHFWURQWKHUPDOKROH $XJHU UHFRPELQDWLRQ SDWKZD\ LQ ORZYROWDJH VKRUWFKDQQHO Q0267 ZLWK HOHFWURQ NLQHWLF HQHUJ\ WKUHVKROG RI H9 ,QWHUIDFH WUDS JHQHUDWLRQ LQ SFKDQQHO 0267 GXULQJ FKDQQHO KRW KROH VWUHVV LV LQYHVWLJDWHG 5HGXFWLRQ RI LQWHUIDFH WUDS GHQVLW\ LV REVHUYHG IRU WKH ILUVW WLPH +\GURJHQ UHOHDVHG IURP WKH K\GURJHQDWHG ERURQ LQ WKH SGUDLQ UHJLRQ E\ WKH FKDQQHO KRW KROH LV WKH SURSRVHG PHFKDQLVP 7KH UHOHDVHG K\GURJHQ GLIIXVHV IURP WKH SGUDLQ YLL

PAGE 8

LQWR WKH FKDQQHO UHJLRQ DQG SDVVLYDWHV WKH IDEULFDWLRQUHVLGXDO LQWHUIDFH WUDSV RYHU WKH FKDQQHO DQG VRXUFH LQWHUIDFH

PAGE 9

&+$37(5 ,1752'8&7,21 5HFHQW DGYDQFHPHQW LQ VLOLFRQ 9HU\/DUJH6FDOH,QWHJUDWLRQ 9/6,f DQG 8OWUD /DUJH6FDOH,QWHJUDWLRQ 8/6,f WHFKQRORJ\ SDUWLFXODUO\ OLWKRJUDSK\ KDV OHG WR DQ DJJUHVVLYH VFDOLQJ RI PHWDOR[LGHVHPLFRQGXFWRU WUDQVLVWRUV 0267Vf WR VPDOOHU GLPHQVLRQV WR LQFUHDVH WKH SDFNLQJ GHQVLW\ WUDQVLVWRUVFKLSf DQG WR LQWHJUDWH PRUH FRPSOH[ FLUFXLW IXQFWLRQV RQWR D VLQJOH VLOLFRQ FKLS RU GLH >@ 7KH VPDOO WUDQVLVWRU VL]H DOVR JUHDWO\ LPSURYHV WKH VSHHG RI WKH 9/6,8/6, FLUFXLWV )RU LQVWDQFH LQ WKH f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f DQG VXSSO\ YROWDJH ZLWK LQFUHDVHG GRSLQJ FRQFHQWUDWLRQ WR VFDOH 0267 WR VPDOOHU IHDWXUHV 7KLV DSSURDFK DWWHPSWV WR PDLQWDLQ D ORQJFKDQQHO FKDUDFWHULVWLF LQ VFDOHG 0267 DQG DOVR WR NHHS WKH

PAGE 10

GUDLQ HOHFWULF ILHOG DQG SRZHU GHQVLW\ FRQVWDQW DIWHU VFDOLQJ +RZHYHU WKH SUHVHQW WUDQVLVWRU VFDOLQJ WUHQG KDV IROORZHG WKH FRQVWDQWYROWDJH VFDOLQJ ODZ PRUH FORVHO\ LQ RUGHU WR EH FRPSDWLEOH ZLWK SRZHU VXSSO\ YROWDJH DQG QRLVH PDUJLQ FRQVWUDLQWV %XW VFDOLQJ GRZQ WKH JDWH R[LGH WKLFNQHVV DQG FKDQQHO OHQJWK DW FRQVWDQW VXSSO\ YROWDJH XQDYRLGHGO\ LQFUHDVHV WKH R[LGH DQG FKDQQHO HOHFWULF ILHOGV 7KHVH KLJK ILHOGV UHGXFH 0267 UHOLDELOLW\ EHFDXVH WKH\ FDXVH f DFFHOHUDWLRQ RI KRW FDUULHUV HOHFWURQV RU KROHVf WR KLJK NLQHWLF HQHUJLHV f LQMHFWLRQ RI KRW FDUULHUV YLD VXUPRXQWLQJ WKH 6L6L HQHUJ\ EDUULHU LQWR WKH JDWH R[LGH f TXDQWXPPHFKDQLFDO EDQGWREDQG DQG EDQGWUDSEDQG WXQQHOLQJ RI WKHUPDO FDUULHUV LQWR WKH JDWH R[LGH f FKDUJLQJ DQG JHQHUDWLRQ RI R[LGH WUDSV E\ WKH LQMHFWHG FDUULHUV DQG f JHQHUDWLRQ RI LQWHUIDFH WUDSV >@ 7KH SK\VLFDO FKDUJLQJGLVFKDUJLQJ DQG JHQHUDWLRQDQQLKLODWLRQ RI WKH R[LGH DQG LQWHUIDFH WUDSV E\ WKH KRW FDUULHUV KDYH EHHQ LGHQWLILHG WR KDYH GHWULPHQWDO HIIHFWV RQ WKH VWDELOLW\ DQG UHOLDELOLW\ RI WKH VLOLFRQ 0267V 7KHVH KRW FDUULHUV HIIHFWV FDXVH D JUDGXDO GHJUDGDWLRQ RI WKH WUDQVLVWRU HOHFWULFDO FKDUDFWHULVWLFV DV UHIOHFWHG E\ FKDQJHV LQ WUDQVLVWRUfV WKUHVKROG YROWDJH $9JWf GUDLQFXUUHQW GULYLQJ FDSDELOLW\ $,'VDWf VXEWKUHVKROG GLVWRUWLRQ $6f DQG DQ LQFUHDVH LQ VWDQGE\ SRZHU FRQVXPSWLRQ :KHQ WKH GHJUDGDWLRQ RI WKH 0267 H[FHHGV DQ DFFHSWDEOH OHYHO LQ D 9/6,8/6, FLUFXLW IRU H[DPSOH $9*7 P9 RU A'VDWA'VDW2 A WKH FLUFXLW PD\ QRW RSHUDWH SURSHUO\ DQG UHOLDEOH RSHUDWLRQ RI WKH HOHFWURQLF V\VWHP XVLQJ WKLV FLUFXLW ZLOO EH DGYHUVHO\ DIIHFWHG 7KHUHIRUH WKH KRW FDUULHU HIIHFWV LQ GHHS VXEPLFURQ 0267 PXVW EH XQGHUVWRRG DQG PLQLPL]HG VXFK WKDW WKH VFDOHGGRZQ 0267V FDQ FRQWLQXH WR DFKLHYH D UHOLDEOH RSHUDWLRQ 7LPHWR)DLOXUH 77)f RI \HDUV ,QQRYDWLYH KRWFDUULHUUHVLVWDQW WUDQVLVWRU GHVLJQV DQG FDUHIXO SURFHVVLQJ

PAGE 11

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f 77) H[WUDFWLRQ PHWKRGRORJ\ WR VWUHVV D %LSRODU -XQFWLRQ 7UDQVLVWRU %-7f DW ORZ HPLWWHUEDVH UHYHUVHELDV YROWDJHV 7KH FXUUHQW DFFHOHUDWLRQ VWUHVV WHFKQLTXH XVHV LQFUHDVHG KRW FDUULHU GHQVLW\ DW ORZ VWUHVV YROWDJHV WR

PAGE 12

DFFHOHUDWH WKH GHJUDGDWLRQ UDWH RI WKH %-7V :H ZLOO GHPRQVWUDWH WKH H[WHQVLRQ RI WKLV FXUUHQWDFFHOHUDWHG %-7 VWUHVV PHWKRGRORJ\ WR 0267 XQGHU FKDQQHO KRW FDUULHUV &+&f VWUHVV )RU 0267 WKH GHQVLW\ RI WKH KRW FDUULHUV LV LQFUHDVHG E\ IRUZDUGELDVLQJ WKH VXEVWUDWHERG\ RU VRXUFHERG\ QS MXQFWLRQ GXULQJ &+& VWUHVV 7KLV DSSURDFK ZLOO EH NQRZQ DV FXUUHQWDFFHOHUDWHG FKDQQHO KRW FDUULHUV VWUHVV &$&+&f 5HVXOWV ZLOO EH DQDO\]HG XVLQJ D SK\VLFEDVHG GHJUDGDWLRQ PRGHO 7KH PHWKRGRORJ\ DGRSWHG LQ WKLV ZRUN ZLOO EH EDVHG RQ 6WUHVV$QG0HDVXUH 6$0f H[SHULPHQWV 6WUHVV UHIHUV WR FRQYHQWLRQDO VXEVWUDWH KRW FDUULHU 6+&f DQG &+& VWUHVVHV RU WKH QHZ &$&+( VWUHVV SURSRVHG DQG GHPRQVWUDWHG LQ WKH IROORZLQJ FKDSWHU 0HDVXUH UHIHUV WR HOHFWULFDO FKDUDFWHUL]DWLRQ RI WKH GHYLFH GHJUDGDWLRQ 7KLV 6$0 PHWKRGRORJ\ ZLOO EH GHVFULEHG LQ &KDSWHU ZKHUH DQ RYHUYLHZ RI VXEVWUDWH DQG FKDQQHO KRW FDUULHUV VWUHVVHV LV SUHVHQWHG &KDSWHU SUHVHQWV D QRYHO GLUHFWFXUUHQW FXUUHQWYROWDJH '&,9f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

PAGE 13

$9*W DQG $6f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f H9 IRU KRWHOHFWURQKRWKROH LQWHUEDQG LPSDFW JHQHUDWLRQ f H9 IRU ERWK KRWHOHFWURQWKHUPDOKROH LPSDFW FROOLVLRQ SDWKZD\ DQG KRWHOHFWURQWKHUPDOKROH H[FKDQJH UHFRPELQDWLRQ SDWKZD\ DQG f H9 IRU KRWHOHFWURQWKHUPDO KROH $XJHU UHFRPELQDWLRQ SDWKZD\ ([SHULPHQWV DUH GHVFULEHG ZKLFK YHULILHG WKH H[LVWHQFH RI WKH H9 WKUHVKROG HQHUJ\ IRU LQWHUEDQG LPSDFW JHQHUDWLRQ SDWKZD\ &KDSWHU SUHVHQWV IRU WKH ILUVW WLPH DQ LQWHUIDFH WUDS UHGXFWLRQ SURFHVV REVHUYHG XVLQJ '&,9 PHDVXUHPHQW LQ SFKDQQHO 0267 GXULQJ FRQYHQWLRQDO FKDQQHOKRWKROH

PAGE 14

VWUHVV $ K\GURJHQGLIIXVLRQ PRGHO ZLOO EH SURSRVHG WR DFFRXQW IRU WKH UHGXFWLRQ RI LQWHUIDFH WUDSV LQ WKH FKDQQHO &KDSWHU JLYHV WKH VXPPDULHV DQG FRQFOXGHV WKLV WKHVLV

PAGE 15

&+$37(5 675(66$1'0($685( 6$0f 0(7+2'2/2*< ,QWURGXFWLRQ ,Q D 6WUHVV$QG0HDVXUH 6$0f H[SHULPHQW WKH GHJUDGDWLRQ FKDUDFWHULVWLFV RI D 026 WUDQVLVWRU 0267f VXEMHFW WR HLWKHU VXEVWUDWH KRWFDUULHU 6+&f RU FKDQQHO KRW FDUULHU &+&f VWUHVV DUH PRQLWRUHG E\ WKH IROORZLQJ HOHFWULFDO PHDVXUHPHQWV f GUDLQ FXUUHQW YV JDWH YROWDJH ,'9*f f VDWXUDWLRQ GUDLQ FXUUHQW YV JDWH YROWDJH ORVDUORff f 'LUHFW&XUUHQW &XUUHQW9ROWDJH '&,9f RU ,%9*% DQG f IRUZDUGELDVHG JDWHG HPLWWHUEDVH SQ MXQFWLRQ ,%9%(f 7KHVH VWUHVV DQG PHDVXUH VWHSV DUH DXWRPDWHG YLD ,((( EXV LQ D 6$0 VWDWLRQ ,Q WKLV FKDSWHU WKH 6$0 VWDWLRQ VHWXS DQG WKH PHDVXUHPHQW DQG VWUHVV ELDV FRQILJXUDWLRQV DUH GHVFULEHG $Q RYHUYLHZ RI VXEVWUDWH DQG FKDQQHOKRW HOHFWURQ VWUHVV LV DOVR SUHVHQWHG 6$0 6HWXS )LJXUH VKRZV D EORFN GLDJUDP RI WKH 6$0 VWDWLRQ VHWXS ,W FRQVLVWV RI WKUHH SDUWV f D 'HYLFH8QGHU7HVW '87f ER[ f DQ LQVWUXPHQWDWLRQ UDFN RI ,((( EXV FRQWUROOHG SRZHU VXSSOLHV YROWDJHFXUUHQW SXOVH JHQHUDWRUV GLJLWDO FXUUHQW DQG YROWn PHWHUV DQG f D 'LJLWDO (TXLSPHQW &RRSRUDWLRQ 0LFUR9D[,, FRPSXWHU 7KH 026 WUDQVLVWRU '87f IDEULFDWHG RQ D RU ZDIHU LV ORDGHG LQWR D OLJKWVKLHOGHG ZDIHU SUREH VWDWLRQ ZKHUH WKH 'UDLQ 'f *DWH *f 6RXUFH 6f %RG\ ; RU %f DQG 6XEVWUDWH 6XE RU ( IRU (PLWWHUf RI WKH '87 DUH FRQQHFWHG YLD WKH SUREHV DQG VKLHOGHG ORZ

PAGE 16

6RXUFH 6fB 'HYLFH *DWH *f ,((( ,QWHUIDFH EXV 'UDLQ 'f ,QVWUXPHQW 8QGHU 7HVW r r 287f -RG\%DVH %f A6XEVWUDWH (A )LJXUH %ORFN GLDJUDP RI WKH 6WUHVVDQG0HDVXUH 6$0 6WDWLRQ '(& 0LFUR9D[ ,, &RPSXWHU

PAGE 17

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t .f DQG +3 PXOWLPHWHUV DUH XVHG WR PHDVXUH WKH GF FXUUHQWV GXULQJ VWUHVV DQG PHDVXUHPHQW F\FOHV VHH )LJXUH f 7KH PHDVXUHPHQW DQG VWUHVV FRQILJXUDWLRQV DUH PXOWLSOH[HG E\ D +3 VFDQQHU DQG D FXVWRPEXLOW 5HHG VZLWFKLQJ UHOD\V ER[ 7KH +3 VFDQQHU FRQWDLQV WZR UHOD\V FDUGV f D FKDQQHO DUUD\ RI VLQJOHSROHVLQJOHWKURZ UHOD\V FRQWUROV WKH 5HHG UHOD\ ER[ DQG f D FKDQQHO DUUD\ RI GRXEOHSROHVLQJOHWKURZ UHOD\V LV PXOWLSOH[HG WRJHWKHU ZLWK WKH +3 YROWPHWHU WR PHDVXUH WKH WHUPLQDO YROWDJHV DSSOLHG RQ '87 (OHFWULFDO 0HDVXUHPHQW 'UDLQ&XUUHQW 0HDVXUHPHQW 7ZR GUDLQ FXUUHQW PHDVXUHPHQW FRQILJXUDWLRQV DUH LPSOHPHQWHG E\ WKH 6$0 VWDWLRQ f OLQHDU ,R9*6 DQG f VDWXUDWLRQ ,SVDWnA'6 7KH GLIIHUHQFH EHWZHHQ WKH OLQHDU DQG VDWXUDWLRQ GUDLQ FXUUHQW PHDVXUHPHQW GHSHQGV RQ KRZ WKH GUDLQ DQG JDWH DUH FRQQHFWHG HOHFWULFDOO\ )LJXUH DQG )LJXUH LOOXVWUDWH WKH VFKHPDWLF RI WKH 6$0 VWDWLRQ VHWXS IRU OLQHDU DQG VDWXUDWLRQ ,' PHDVXUHPHQW UHVSHFWLYHO\ 'XULQJ WKH OLQHDU

PAGE 18

)LJXUH )ORZ FKDUW RI WKH 6$0 FRQWURO SURJUDP

PAGE 19

. )LJXUH 6$0 VFKHPDWLF VHWXS IRU OLQHDU ,'9* PHDVXUHPHQW

PAGE 20

+3 )LJXUH 6$0 VFKHPDWLF VHWXS IRU VDWXUDWLRQ ,'VDW PHDVXUHPHQW

PAGE 21

DQG VDWXUDWLRQ ,' PHDVXUHPHQW WKH VRXUFH ERG\ RU SZHOOf DQG WKH VXEVWUDWH RI WKH 0267 DUH JURXQGHG )RU WKH OLQHDU ,'9* PHDVXUHPHQW D SURJUDPPDEOH YROWDJH LV VXSSOLHG E\ WR JLYH D ORZ 9f DQG FRQVWDQW YROWDJH DFURVV WKH GUDLQ DQG VRXUFH 9'6 DQG DQRWKHU YROWDJH VRXUFH LV FRQQHFWHG WR WKH JDWH WR VWHS WKH JDWH YROWDJH LQ VPDOO LQFUHPHQWDO VWHS IURP DFFXPXODWLRQ WR LQYHUVLRQ 7KH GUDLQ FXUUHQW LV PHDVXUHG E\ D VHQVLWLYH SLFRDPPHWHU .f 7KH YROWDJHV DSSOLHG WR WKH JDWH DQG GUDLQ WHUPLQDOV DUH PHDVXUHG E\ WKH +3 GLJLWDO YROWPHWHU 7KXV WKH ,'9* FKDUDFWHULVWLFV RI WKH 0267 LV REWDLQHG E\ PHDVXULQJ WKH GUDLQ FXUUHQW IRU D UDQJH RI JDWH YROWDJHV 'XULQJ WKH VDWXUDWLRQ ,'VDW9' PHDVXUHPHQW WKH JDWH DQG GUDLQ DUH WLHG WRJHWKHU DQG FRQQHFWHG WR D VLQJOH YROWDJH VRXUFH VHH )LJXUH f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f FXUUHQW YHUVXV JDWH YROWDJH ZLWK D IRUZDUGELDVHG SQ MXQFWLRQ VXSSO\LQJ WKH PLQRULW\ FDUULHUV WR UHFRPELQH ZLWK WKH PDMRULW\ FDUULHUV DW UHFRPELQDWLRQ FHQWHUV DW WKH 6L6L LQWHUIDFH RI WKH 0267 ,Q WKLV VHFWLRQ ZH ZLOO RQO\ GHVFULEH WKH EDVLF PHDVXUHPHQW

PAGE 22

. )LJXUH 6$0 VFKHPDWLF VHWXS IRU '&,9 PHDVXUHPHQW VKRZLQJ WKH 'UDLQHPLWWHU '&,9 FRQILJXUDWLRQ

PAGE 23

SURFHGXUH 'HWDLOV RI '&,9 PHDVXUHPHQW FRQILJXUDWLRQV DQG LWV EDVLF SULQFLSOHV ZLOO EH HODERUDWHG LQ &KDSWHU )LJXUH VKRZV WKH VFKHPDWLF VHWXS IRU 'UDLQ(PLWWHU '&,9 PHDVXUHPHQW $ YROWDJH VRXUFH FRQVWDQW 9(%f LV FRQQHFWHG WR WKH GUDLQ WR IRUZDUG ELDV WKH GUDLQ QS MXQFWLRQ $QRWKHU YROWDJH VRXUFH 9*%f LV FRQQHFWHG WR WKH JDWH WR PRGXODWH WKH 6L VXUIDFH IURP DFFXPXODWLRQ WR LQYHUVLRQ $ SLFRDPPHWHU LV FRQQHFWHG WR WKH SZHOO WR PHDVXUH WKH ERG\ FXUUHQW DW HDFK 9*% VWHS 7KH VRXUFH LV JURXQGHG GXULQJ WKH HQWLUH PHDVXUHPHQW *DWHG ,%9%% 0HDVXUHPHQW 7KH ,%9%( PHDVXUHPHQW LV LGHQWLFDO WR WKH VWDQGDUG PHWKRG XVHG IRU ELSRODU WUDQVLVWRUVf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f FXUUHQW LV PRQLWRUHG XVLQJ WKH SLFRDPPHWHU 6XEVWUDWH DQG &KDQQHO +RW (OHFWURQ 6WUHVV ,Q VLOLFRQ WKHUPDO HOHFWURQV DW WKH FRQGXFWLRQ EDQG HGJH FDQ JDLQ NLQHWLF HQHUJ\ E\ WUDQYHUVLQJ WKURXJK D SRWHQWLDO GURS DQG EHFRPHV HQHUJHWLF DQG KRW 7KHVH KRW HOHFWURQV DUH FKDUDFWHUL]HG E\ DQ HIIHFWLYH WHPSHUDWXUH 7HII fm.LQHWLF (QHUJ\fN% ZKLFK LV PXFK JUHDWHU WKDQ WKH ODWWLFH WHPSHUDWXUH 7/DWWLFH )RU H[DPSOH LI

PAGE 24

.( H9 7HII f.OH9H9f } 7ODWWLFH DW URRP WHPSHUDWXUH ,Q Q0267 WKHUH DUH WZR LPSRUWDQW KRWHOHFWURQ HIIHFWV f WKH VXEVWUDWH KRWHOHFWURQ 6+(f HIIHFW ZKHUH WKH SULPDU\ VRXUFH RI KRW HOHFWURQV LV GXH WR WKH FXUUHQW IURP WKH VXEVWUDWH DQG f WKH FKDQQHO KRWHOHFWURQ &+(f HIIHFW ZKHUH WKH SULPDU\ VRXUFH RI KRW HOHFWURQV LV GXH WR WKH VXUIDFH FKDQQHO FXUUHQW 7R LQYHVWLJDWH WKHVH WZR KRWHOHFWURQ HIIHFWV LQ VLOLFRQ 0267 WKH VWUHVV PRGH LPSOHPHQWHG RQ WKH 6$0 VWDWLRQ LV FDUHIXOO\ GHVLJQHG E\ FRQVLGHULQJ WKH GHVLUHG VWUHVV FRQGLWLRQV DQG FRQILJXUDWLRQV 7KHVH ZLOO EH GHVFULEHG LQ WKH IROORZLQJ WZR VHFWLRQV 6XEVWUDWH +RW (OHFWURQ 6WUHVV ,Q WKLV ZRUN D VXEVWUDWH KRW HOHFWURQ LQMHFWLRQ 6+(Lf WHFKQLTXH LV XVHG WR XQLIRUPO\ LQMHFW HOHFWURQV IURP WKH VLOLFRQ VXEVWUDWH LQWR WKH JDWH R[LGH RI WKH 0267 7KLV LQMHFWLRQ WHFKQLTXH HPSOR\V D IRUZDUGELDVHG YHUWLFDO ELSRODU WUDQVLVWRU >@ DQG KDV EHHQ SUHYLRXVO\ XVHG WR HPSOR\HG WR VWXG\ WKH EDVLF GHJUDGDWLRQ PHFKDQLVPV RI WKH 6L JDWH R[LGH DQG LWV IXQGDPHQWDO HOHFWULFDO SURSHUWLHV > @ 2SWLFDO KRW HOHFWURQ LQMHFWLRQ 2+(Lf >@ LV D YDULDWLRQ RI 6+(L 7KH PDMRU GLIIHUHQFH EHWZHHQ 2+(L DQG %L0266+(L LV WKH VRXUFH RI PLQRULW\ FDUULHUV HOHFWURQVf LQ WKH SW\SH VLOLFRQ ERG\ RI WKH Q0267 )RU 2+(L WKH KRW HOHFWURQV FRPHV IURP LQWHUEDQG RSWLFDO JHQHUDWLRQ RI HOHFWURQKROH SDLU LQ WKH VXUIDFH VSDFHFKDUJHUHJLRQ LQ WKH SVLOLFRQ VXEVWUDWH XQGHUQHDWK WKH JDWH R[LGH )RU %L0266+(L WKH VRXUFH RI KRW HOHFWURQV LV WKH PLQRULW\ FDUULHUV HOHFWURQVf LQMHFWHG E\ WKH IRUZDUGELDVHG QS HPLWWHUEDVH MXQFWLRQ %L0266+(L LV HPSOR\HG LQ WKLV ZRUN EHFDXVH RI LWV HDV\ DQG VWUDLJKWIRUZDUG LPSOHPHQWDWLRQ )RU 2+(L D VSHFLDO WHVW VWUXFWXUH ZLWK WKLQ OD\HU RI

PAGE 25

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f DQG WKH SZHOO EDVHf WR ELDV WKH YHUWLFDO SDUDVLWLF ELSRODU WUDQVLVWRU LQWR DFWLYH UHJLRQ DV ZHOO DV WR FRQWURO WKH DPRXQW RI WKH EDQG EHQGLQJ LQ WKH VLOLFRQ VXUIDFH VSDFHFKDUJH OD\HU $ IRUZDUGELDV LV DSSOLHG EHWZHHQ WKH SZHOO EDVHf DQG WKH VXEVWUDWH HPLWWHUf WR VXSSO\ WKH PLQRULW\ FDUULHUV HOHFWURQVf $ ONe UHVLVWRU ZDV FRQQHFWHG LQ VHULHV ZLWK WKH HPLWWHU WR OLPLW WKH HPLWWHU FXUUHQW 7KH HPLWWHU DQG FROOHFWRU FXUUHQWV ZHUH PHDVXUHG E\ WZR SURJUDPPDEOH +3$ DPPHWHUV :KLOH NHHSLQJ WKH UHYHUVHELDV EHWZHHQ WKH SZHOO EDVHf DQG VRXUFHGUDLQ FROOHFWRUf FRQVWDQW WKH IRUZDUG ELDV DFURVV WKH SZHOO EDVHf DQG VXEVWUDWH HPLWWHUf ZDV DGMXVWHG E\ D VRIWZDUH IHHGEDFN ORRS LQ WKH 6$0 FRQWURO SURJUDP WR PDLQWDLQ D FRQVWDQW JDWH LQMHFWLRQ FXUUHQW 7KH PDJQLWXGH RI WKH JDWH

PAGE 26

. )LJXUH 6$0 VFKHPDWLF VHWXS IRU %L0266+( LQMHFWLRQ VWUHVV

PAGE 27

LQMHFWLRQ FXUUHQW FDQ EH FRQWUROOHG E\ YDU\LQJ WKH IRUZDUGELDV DFURVV WKH HPLWWHUEDVH MXQFWLRQ DQG WKH UHYHUVHELDV DFURVV WKH FROOHFWRUEDVH MXQFWLRQ &KDQQHO +RW (OHFWURQ 6WUHVV 7KH FKDQQHO KRW HOHFWURQ VWUHVV FRQGLWLRQV LQ Q0267 DUH FODVVLILHG EDVHG RQ WKH PDJQLWXGH RI WKH DSSOLHG 9'6 DQG 9*6 7ZR FRPPRQO\ XVHG &+( VWUHVV FRQGLWLRQV DUH f 9'6 9*6 RU WKH WULRGH UDQJH DQG f 9'6 9*6 RU GUDLQ FXUUHQWVDWXUDWLRQ UDQJH :KHQ 9'6 9*6 WKH WUDQVLVWRU LV RSHUDWLQJ LQ WKH OLQHDU UDQJH 7KH FKDQQHO HOHFWURQV DUH DFFHOHUDWHG E\ WKH ODUJH FKDQQHO HOHFWULF ILHOG LQ SDUDOOHO ZLWK WKH 6L6L LQWHUIDFH QHDU WKH GUDLQ WR JDLQ PRUH NLQHWLF HQHUJ\ 1HDU WKH GUDLQ MXQFWLRQ WKH HOHFWURQ WUDMHFWRU\ PD\ EH GHIOHFWHG VLGHZD\V WRZDUGV WKH 6L6L LQWHUIDFH E\ SKRQRQ VFDWWHULQJ ,I WKH NLQHWLF HQHUJ\ RI WKHVH LPSLQJLQJ HOHFWURQV LV JUHDWHU WKDQ WKH HOHFWURQ EDUULHU KHLJKW DW WKH 6L6L LQWHUIDFH aH9f WKH\ FDQ EH LQMHFWHG LQWR WKH JDWH R[LGH 7KLV HOHFWURQ JDWH FXUUHQW ZDV PRGHOHG XVLQJ WKH 6KRFNOH\fV OXFN\ HOHFWURQ PRGHO >@ DQG H[SUHVVHG >@ DV ,FKH;3 f ZKHUH ,FK LV WKH FKDQQHO FXUUHQW (0 LV WKH PD[LPXP HOHFWULF ILHOG SHUSHQGLFXODU WR WKH 6L6L LQWHUIDFH DQG LV WKH HPSLULFDO HOHFWURQ VFDWWHULQJ PHDQIUHHSDWK $f 7KH HOHFWURQ EDUULHU KHLJKW Mf%1 LV H9 1LQJ HPSOR\HG DQ HIIHFWLYH HOHFWURQ EDUULHU KHLJKW JLYHQ E\ >@ S%1 [(R[ (R[ f ZKHUH (R[ LV WKH R[LGH HOHFWULF ILHOG QHDU WKH GUDLQ MXQFWLRQ 7KH ODVW WZR WHUPV WDNHV LQWR DFFRXQW RI WKH 6FKRWWN\EDUULHU ORZHULQJ DQG WKH SRVVLELOLW\ RI WXQQHOLQJ WKURXJK

PAGE 28

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f DQG ZLWK DSSUHFLDEOH JDWH FXUUHQW KLJK HOHFWURQ LQMHFWLRQ HIILFLHQF\f ZLOO GHJUDGH IDVWHU XQGHU VDWXUDWLRQSRLQW &+( VWUHVV >@ :KHQ 9'6!9*6 WKH WUDQVLVWRU LV ELDVHG GHHS LQWR WKH GUDLQFXUUHQW VDWXUDWLRQ UDQJH 8QOLNH WKH VDWXUDWLRQSRLQW &+( VWUHVV WKH FKDQQHO HOHFWURQV DUH PDLQO\ DFFHOHUDWHG E\ WKH ODUJH HOHFWULF ILHOG LQ WKH GUDLQ VSDFHFKDUJH OD\HU GXH WR 9' 9'VDW LQVWHDG RI E\ WKH VPDOO FKDQQHO HOHFWULF ILHOG ,I WKH NLQHWLF HQHUJ\ RI WKH KRW HOHFWURQV LV JUHDWHU WKDQ a(*6L LH 9' 9'VDW ,(*B6LTO LQWHUEDQG LPSDFW JHQHUDWLRQ RI HOHFWURQ KROH SDLUV RFFXUV 0RVW RI WKH KROHV IORZ LQWR WKH SZHOO JLYLQJ ULVH WR D ERG\ RU ZHOOf FXUUHQW LQ WKH SZHOO WHUPLQDO ,[f 7KH REVHUYHG ,[ W\SLFDOO\ SHDNV DW 9'6 9*6 > @ $ SHDN ZHOO FXUUHQW ZLOO FRUUHVSRQG WR PD[LPL]LQJ WKH FRPELQDWLRQ RI LQWHUEDQG LPSDFW JHQHUDWLRQ UDWH RI HOHFWURQKROH SDLUV DQG WKH HIILFLHQW FROOHFWLRQ RI KROHV E\ WKH SZHOO 7KLV ZHOOFXUUHQW FDQ RYHUORDG WKH VXEVWUDWHELDV JHQHUDWRU GXULQJ

PAGE 29

RSHUDWLRQ DQG FDXVH D VQDSEDFN EUHDNGRZQ DQG &026 ODWFKXS 'XH WR WKH GLPHQVLRQDO HOHFWULF ILHOG GLVWULEXWLRQ LQ WKH GUDLQ VSDFHFKDUJH OD\HU HPSLULFDO PRGHOV ZLWK DGMXVWDEOH SDUDPHWHUV KDYH EHHQ XVHG WR PRGHO ,[ E\ UHODWLQJ WKH SHDN HOHFWULF ILHOG (Pf DW WKH GUDLQ WR WKH GUDLQ YROWDJH 9'f > @ ,[ $s f ,FK f H[S f§%(Pf f ZKHUH $ DQG % DUH HPSLULFDO ILWWLQJ SDUDPHWHUV 7KH SHDN ZHOO FXUUHQW ZDV DVVXPHG WR KDYH DQ H[SRQHQWLDO GHSHQGHQW RQ WKH PD[LPXP HOHFWULF ILHOG DW WKH GUDLQ HGJH $Q HVWLPDWH RI WKH PD[LPXP FKDQQHO HOHFWULF ILHOG ZDV JLYHQ E\ >@ Y' Y'VDW (P f A f§ < GHS ZKHUH 9'VDW LV WKH GUDLQ VDWXUDWLRQ YROWDJH / LV WKH FKDQQHO OHQJWK DQG \GHS LV FKDQQHO GHSOHWLRQ SRLQW (TQf DVVXPHV WKDW WKH PD[LPXP ORQJLWXGLQDO HOHFWULF ILHOG LQ WKH VLOLFRQ VXUIDFH OD\HU EHWZHHQ WKH VRXUFH DQG GUDLQ RFFXUV EHWZHHQ WKH GHSOHWLRQ SRLQW \ \GHS DQG WKH GUDLQ PHWDOOXUJLFDO MXQFWLRQ \ / 7KH GLVWDQFH IURP WKH FKDQQHO GHSOHWLRQ SRLQW WR WKH GUDLQ PHWDOOXUJLFDO MXQFWLRQ KDV EHHQ HPSLULFDOO\ PRGHOHG E\ &KDQ3R+X >@ / \GHS [R[;M f ZKHUH [; LV WKH JDWH R[LGH WKLFNQHVV DQG ;M LV WKH GUDLQ MXQFWLRQ GHSWK $W ODUJH 9'6 DQG 9*6 9*7f HQHUJHWLF LPSDFWJHQHUDWHG HOHFWURQV DQG KROHV FDQ EH LQMHFWHG VLGHZD\ LQWR WKH JDWH R[LGH >@ %HWZHHQ WKH GHSOHWLRQ SRLQW DQG WKH VRXUFH WKH VXUIDFH DQG R[LGH HOHFWULF ILHOG DWWUDFWV WKH FKDQQHO HOHFWURQV WR WKH 6L6L LQWHUIDFH 7KH VXUIDFH DQG R[LGH HOHFWULF ILHOG LV GLUHFWHG LQ WKH RSSRVLWH GLUHFWLRQ

PAGE 30

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

PAGE 31

)LJXUH VKRZV WKH VHFRQG FXUUHQWDFFHOHUDWHG ZKLFK IRUZDUGELDVHV WKH QVXEVWUDWHSERG\ MXQFWLRQ ,Q WKLV FDVH D QHJDWLYH YROWDJH LV DSSOLHG WR WKH Q VXEVWUDWH RI WKH QFKDQQHO 0267 DQG WKH SERG\ DQG QVRXUFH DUH JURXQGHG GXULQJ &$&+( VWUHVV

PAGE 32

%RG\ 6RXUFH L Q ?“ ? Q 3ZHOO 1HSLWD[LDO OD\HU 1 VXEVWUDWH +3 *DWH , f§_>K3@ 9 n UIn9WI /f f§f§ f§ Y ? “r R 6XEVWUDWH +3 7 +3 )LJXUH 6$0 VFKHPDWLF VHWXS IRU FXUUHQWDFFHOHUDWHG &+( VWUHVV XVLQJ WKH IRUZDUGELDVHG VRXUFHERG\ MXQFWLRQ

PAGE 33

. )LJXUH 6$0 VFKHPDWLF VHWXS IRU FXUUHQWDFFHOHUDWHG &+( VWUHVV XVLQJ WKH IRUZDUGELDVHG VXEVWUDWHERG\ MXQFWLRQ

PAGE 34

&+$37(5 ',5(&7&855(17 &855(1792/7$*( '&,9f &+$5$&7(5,=$7,21 7(&+1,48( ,QWURGXFWLRQ ,Q WKH 0267 ZLWK D QS MXQFWLRQ LVRODWLRQ ZHOO XVXDOO\ WKH S0267 RI WKH SUHVHQW SURGXFWLRQ &026 FKLSVf WKH YHUWLFDO RU ODWHUDO %-7 DFWLRQ LV DWWDLQHG E\ IRUZDUGn ELDVLQJ HLWKHU WKH QS ZHOO MXQFWLRQ RU WKH VRXUFHZHOO RU GUDLQZHOO SQ MXQFWLRQ ,Q WKH 0267 ZLWKRXW WKH QS MXQFWLRQ LVRODWLRQ ZHOO XVXDOO\ WKH Q0267 RI WKH SUHVHQW SURGXFWLRQ &026 FKLSVf RQO\ WKH ODWHUDO %-7 DFWLRQ FDQ EH DWWDLQHG E\ IRUZDUGELDVLQJ WKH GUDLQERG\ RU VRXUFHERG\ SQ MXQFWLRQ ,Q WKLV ZRUN D VLPSOH DQG KLJK VHQVLWLYLW\ 'LUHFW&XUUHQW &XUUHQW9ROWDJH '&,9f WHFKQLTXH LV HPSOR\HG WR FKDUDFWHUL]H WKH KRWFDUULHULQGXFHG GHJUDGDWLRQ LQ VLOLFRQ PHWDOR[LGHVHPLFRQGXFWRU WUDQVLVWRUV 0267Vf ,Q WKLV FKDSWHU WKH EDVLF PHDVXUHPHQW WHFKQLTXH LV GLVFXVVHG 7KLV '&,9 WHFKQLTXH LV EDVHG RQ WKH JDWH YROWDJH FRQWUROOHG HOHFWURQKROH UHFRPELQDWLRQ UDWH DW WKH 6L6L LQWHUIDFH WUDSV 7KLV WHFKQLTXH ZDV ILUVW HPSOR\HG E\ 6DK LQ WR VWXG\ WKH VXUIDFH UHFRPELQDWLRQ DQG FKDQQHO RQ VLOLFRQ WUDQVLVWRUV >@ 5HFHQWO\ WKLV WHFKQLTXH KDV EHHQ UHLQWURGXFHG E\ 6DK DQG 1HXJURVFKHO WR VWXG\ WKH HIIHFW RI WKH VWUHVVFUHDWHG R[LGH 127f DQG LQWHUIDFH 1,7f WUDSV RQ WKH UHOLDELOLW\ RI DGYDQFHG 026 DQG %LSRODU MXQFWLRQ WUDQVLVWRUV 0267V DQG %-7Vf > @

PAGE 35

0HDVXUHPHQW &RQILJXUDWLRQV 7KH '&,9 WHFKQLTXH LV D JDWHFRQWUROOHG %-7 PHDVXUHPHQW DQG HPSOR\V WKH ODWHUDO RU YHUWLFDO %-7 RI WKH VLOLFRQ 0267 VWUXFWXUH 7KLV ZDV RQFH UHIHUUHG WR DV %L0267 >@ ZKLFK ZLOO EH GLVFDUGHG DQG UHSODFHG E\ WKH DERYH JHQHUDO GHVFULSWLRQ > @ )LJXUH VKRZV D FURVVVHFWLRQDO YLHZ RI D Q0267 LOOXVWUDWLQJ WKH YHUWLFDO DQG ODWHUDO %-7V )LJXUH LOOXVWUDWHV WKUHH QS MXQFWLRQV WKDW FDQ EH IRUZDUGELDVHG WKH HPLWWHU WR LQMHFW PLQRULW\ FDUULHUV LQWR WKH SEDVH ZHOO f WKH 'UDLQ%DVH %RG\f QS MXQFWLRQ f WKH 6RXUFH%DVH %RG\f QS MXQFWLRQ DQG f WKH 6XEVWUDWH%DVH ERG\f QS MXQFWLRQ ,Q WKLV ZRUN WKH ERG\ DQG EDVH DUH XVHG LQWHUFKDQJHDEO\ LQ WKH GLVFXVVLRQ ZLWK WKH VDPH VXEVFULSW % RU E 7KXV WKHUH DUH WKUHH EDVLF '&,9 PHDVXUHPHQW FRQILJXUDWLRQV WZR WRSHPLWWHU f 'UDLQ(PLWWHUV'( RU f 6RXUFH(PLWWHU 6( DQG f ERWWRPHPLWWHUV%( 6XEVWUDWH(PLWWHU LV QRW XVHG WR DYRLG FRQIXVLRQ ZLWK 6RXUFH (PLWWHU DEEUHYLDWLRQf ,Q WKH '('&,9 FRQILJXUDWLRQ WKH QGUDLQSERG\ LV WKH IRUZDUGELDVHG HPLWWHU MXQFWLRQ DQG WKH QVRXUFHSERG\ LV WKH VKRUWFLUFXLWHG QFROOHFWRUSEDVH MXQFWLRQ 6LPLODUO\ LQ WKH 6('&,9 FRQILJXUDWLRQ WKH QVRXUFHSERG\ MXQFWLRQ LV WKH IRUZDUGELDVHG HPLWWHU MXQFWLRQ DQG WKH QGUDLQSERG\ LV WKH VKRUWFLUFXLWHG QFROOHFWRUSEDVH MXQFWLRQ ,Q WKH %('&,9 FRQILJXUDWLRQ WKH QVXEVWUDWHSERG\ LV WKH IRUZDUGELDVHG HPLWWHU MXQFWLRQ DQG WKH QGUDLQSERG\ DQGRU QVRXUFHSERG\ LVDUH WKH VKRUWFLUFXLWHG QFROOHFWRUSEDVH MXQFWLRQVf 7KLV V\VWHPDWLF FODVVLILFDWLRQ DFURQ\QHG E\ 6DK > @ RI WKH '&,9 PHDVXUHPHQW FRQILJXUDWLRQV LV QHFHVVDU\ WR H[HPSOLI\ WKH SK\VLFDO ORFDWLRQ RI WKH &+&LQGXFHG GHJUDGDWLRQ DV ZLOO EH GLVFXVVHG ODWHU

PAGE 36

6RXUFH *DWH 'UDLQ %DVH 6Lr 4W R4R7 $ $ $BB $ $ ‘‘ O Q Q SEDVH FRQWDFW £ L SEDVH ZHOO QHSLWD[LDO OD\HU Q VXEVWUDWH 6XEVWUDWH )LJXUH &URVVVHFWLRQDO YLHZ RI D QFKDQQHO 0267 VKRZLQJ WKH ODWHUDO DQG YHUWLFDO SDUDVLWLF QSQ %-7

PAGE 37

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f DORQJ WKH 6L6L LQWHUIDFH FDQ EH JHQHUDWHG RYHU WKH GUDLQERG\ MXQFWLRQ VSDFHFKDUJHUHJLRQ DQG RYHU WKH QFKDQQHO UHJLRQ IURP KRW FDUULHUV EUHDNLQJ WKH VWUDLQHG RU ZHDN K\GURJHQ ERQGV 6L+ DQG 6L2+f 7KXV GXULQJ '('&,9 PHDVXUHPHQW WKH PHDVXUHG ,%B'( FRQWDLQV WZR HPLWWHUV QRW VKRZQ LQ )LJXUH f DQG VL[ EDVH UHFRPELQDWLRQ FRPSRQHQWV > @ 7KH WZR HPLWWHU FRPSRQHQWV DUH WKH HOHFWURQKROH HKf UHFRPELQDWLRQV DW WKH UHFRPELQDWLRQ FHQWHUV LQ WKH GUDLQfV EXONTXDVLQHXWUDO UHJLRQ ,JHLEXLNf DU!G DW WKH LQWHUIDFH EHWZHHQ SRO\VLOLFRQ DQG FU\VWDOOLQHVLOLFRQ FRQWDFWV ,%HLFRQWDFWf! UHVSHFWLYHO\ 7KH EDVH FRPSRQHQWV DUH GXH WR HK UHFRPELQDWLRQV FHQWHUV RU LQWHUIDFH WUDSV ORFDWHG LQ WKH VL[ UHJLRQV ,Q )LJXUH WKHVH VL[ EDVH UHFRPELQDWLRQ FRPSRQHQWV DUH GLVWLQJXLVKHG E\ RQH DOSKDEHWLF DQG RQH QXPHULF VXEVFULSW FAFKDQQHO UHJLRQ VVVXUIDFH VSDFHFKDUJH UHJLRQ E EXON VSDFHFKDUJH UHJLRQ DQG IURP ,% m H[ST9%(QN%7f ZKHUH Q O IRU 6KRFNOH\fV H[DFW SQ MXQFWLRQ GLRGH ODZ DQG Q IRU WKH VLPSOLILHG 6DK1R\FH6KRFNOH\ 616f VSDFHFKDUJHOD\HU UHFRPELQDWLRQ ODZ > @ 7KHVH VL[ FRPSRQHQWV DUH f ,EBGHF RYHU QFKDQQHOLQWHUIDFH f ,EBGHV RYHU

PAGE 38

S6XEVWUDWH )LJXUH &URVVVHFWLRQDO YLHZ RI QFKDQQHO 0267 ELDVHG LQ WKH 'UDLQ(PLWWHU '(f '&,9 FRQILJXUDWLRQ 6L[ EDVH FXUUHQW UHFRPELQDWLRQ SDWKZD\V DUH GHSLFWHG :KLWH VTXDUHV DQG KH[DJRQV LQ 6L DUH R[LGH HOHFWURQ DQG KROH WUDSV :KLWH DQG EODFN WULDQJOHV DUH DFWLYH 6Lm DQG 6L2}f DQG LQDFWLYH K\GURJHQDWHG 6L+ DQG 6L2+f LQWHUIDFH WUDSV 6TXDUHJULGVKDGHG DUHDV DUH VSDFHFKDUJH UHJLRQV RI SQ DQG SS MXQFWLRQV LQ 6L %ODFN DQG ZKLWH FLUFOHV DUH HOHFWURQ DQG KROHV LQ 6L :KLWH VTXDUHV LQ 6L DUH HOHFWURQKROH UHFRPELQDWLRQJHQHUDWLRQ FHQWHUV $GDSWHG IURP 6DK >@

PAGE 39

WKH VXUIDFH VSDFHFKDUJH UHJLRQ RI WKH GUDLQERG\ MXQFWLRQ DORQJ WKH QFKDQQHO SHULPHWHU f ,%B'(E rQ WKH GUDLQERG\ MXQFWLRQ EXON VSDFHFKDUJH UHJLRQ f ,%B'(EO LQ WKH GUDLQERG\ MXQFWLRQ EXONTXDVLQHXWUDO UHJLRQ f ,%B'(Eff VDPH DV f EXW RXWVLGH RI WKH QFKDQQHOSHULPHWHU DQG f ,EGHF,ff VDPH DV f EXW RXWVLGH RI WKH Q FKDQQHOSHULPHWHU 7KXV WKH PHDVXUHG ,%'( LV D VXP RI WKH FRPSRQHQWV GHVFULEHG DERYH A%'( f§ A%HOEXON A%HOFRQWDFW A%'(FO ,%'(V ,%'(E A%'(EO %'(Af A%'(FOf f 2I DOO WKH EDVHFXUUHQW FRPSRQHQWV GHVFULEHG LQ f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f 'XULQJ %('&,9 PHDVXUHPHQW WKH FROOHFWRUEDVH MXQFWLRQ LV HLWKHU JURXQGHG RU ZHDNO\ UHYHUVHELDVHG 7KXV WKH HOHFWURQKROH UHFRPELQDWLRQ FRPSRQHQWV DORQJ WKH VXUIDFH VSDFHFKDUJHUHJLRQ DQG LQ WKH EXON VSDFHFKDUJH UHJLRQ RI WKH FROOHFWRQIRDVH GUDLQERG\ DQG VRXUFHERG\f MXQFWLRQV DUH QHJOLJLEOH GXH WR ORZ HOHFWURQ DQG KROH FRQFHQWUDWLRQV LQ WKHVH UHJLRQV ,Q IDFW WKHUPDO JHQHUDWLRQ RI HOHFWURQKROH

PAGE 40

%RG\ %DVH 9 *% *DWH 'UDLQ &ROOHFWRU OF 1RJMOHF 1R7KRLM JM4 n 1LW 1_7+ $ ADDDDD, 6L2" S%RG\ S:HOO L PUL7Q7UUP ,6KV Q6XEVWUDWH f,% [! PPWW Q -! ;AA fO%AIIJ f 7)57 V 9APY, AEVWUDWH %( (PLWWHU )LJXUH &URVVVHFWLRQDO YLHZ RI QFKDQQHO 0267 ELDVHG LQ WKH %RWWRP(PLWWHU %(f '&,9 FRQILJXUDWLRQ 7KUHH EDVH FXUUHQW UHFRPELQDWLRQ SDWKZD\V DUH GHSLFWHG 6HH )LJXUH FDSWLRQ IRU WKH GHVFULSWLRQ RI V\PEROV

PAGE 41

SDLUV SUHYDLO LQ WKHVH UHJLRQV EXW FDQ EH QHJOHFWHG EHFDXVH WKH ODUJH EDVH UHFRPELQDWLRQ FXUUHQW ZLOO RYHUZKHOP WKLV VPDOO JHQHUDWLRQ FXUUHQW HYHQ DW VPDOO IRUZDUG ELDV >@ 7KXV WKH PHDVXUHG ,% LQ WKH %( FRQILJXUDWLRQ FRQVLVWV RI RQO\ WZR HPLWWHU FRPSRQHQWV DQG WKUHH EDVH UHFRPELQDWLRQ FRPSRQHQWV 7KH WZR HPLWWHU FRPSRQHQWV DUH WKH HK UHFRPELQDWLRQV DW WKH UHFRPELQDWLRQ FHQWHUV LQ WKH QVXEVWUDWHfV EXONTXDVLQHXWUDO UHJLRQ L%%(HOEXONf DUG DW WKH LQWHUIDFH EHWZHHQ WKH Q VXEVWUDWH DQG PHWDO FRQWDFWV A%%(HLFRQWDFWff UHVSHFWLYHO\ 7KH WKUHH EDVH HK UHFRPELQDWLRQ FRPSRQHQWV DUH f ,%EHF RYHU WKH QFKDQQHOLQWHUIDFH ZKLFK LV VDPH DV ,EGHFf f O%%(E LQ WKH QVXEVWUDWHERG\ MXQFWLRQ EXON VSDFHFKDUJH UHJLRQ f ,EEHFW VDPH DV f EXW RXWVLGH RI WKH QFKDQQHOSHULPHWHU 7KXV WKH PHDVXUHG ,%%( ,6 JLYHQ E\ ,E%( A%%(HOEXON A%%(HOFRQWDFW A%%(FO A%(EOf A%%(FOf f ,Q f RQO\ ,EEHF LV PRGXODWHG E\ WKH JDWH YROWDJH 7KXV WKH %('&,9 PHDVXUHPHQW FRQILJXUDWLRQ JLYHV JUHDWO\ VLPSOLILHG GDWD DQDO\VLV DQG SURYLGHV WKH XQLTXH ORFDWLRQLQJ RI WKH 1>7 Q Of LQ WKH FKDQQHO UHJLRQ 7KLV %( FRQILJXUDWLRQ KDG EHHQ XVHG E\ RWKHU UHVHDUFKHUV WR VWXG\ WKH VXUIDFH UHFRPELQDWLRQ UDWH RU YHORFLW\ 6 DW WKH 6L6L LQWHUIDFH > @ DQG WKH DUHDO XQLIRUPLW\ RI SRVLWLYH R[LGH FKDUJH EXLOGXS DQG LQWHUIDFH WUDS JHQHUDWLRQ XVLQJ VXEVWUDWH KRWHOHFWURQ LQMHFWLRQ >@ 7KHRU\ RI 6XUIDFH 5HFRPELQDWLRQ 5DWH 7KH VXUIDFH UHFRPELQDWLRQ FXUUHQW FRPSRQHQWV GLVFXVVHG LQ VHFWLRQ FDQ EH HYDOXDWHG E\ H[WHQGLQJ WKH 6KRFNOH\5HDG+DOO 65+f UHFRPELQDWLRQ NLQHWLF IRU D VLQJOH WUDS OHYHO >@ WR GHWHUPLQH WKH VXUIDFH UHFRPELQDWLRQ UDWH 5VV RI HOHFWURQ DQG

PAGE 42

KROH WKURXJK WKH LQWHUIDFH WUDS ZKLFK KDV D GLVFUHWH HQHUJ\ OHYHO (M LQ WKH VLOLFRQ HQHUJ\ JDS 7KH IROORZLQJ SUHVHQWDWLRQ LV EDVHG RQ WKH OHFWXUHV JLYHQ E\ 6DK >@ &RQVLGHU DQ LQILQLWHVLPDO LQWHUIDFLDO OD\HU RI WKLFNQHVV [ DW WKH 6L6L LQWHUIDFH ZKHUH WKH HOHFWURQKROH UHFRPELQDWLRQ RFFXUV )LJXUH f WKH VWHDG\VWDWH HOHFWURQKROH UHFRPELQDWLRQ UDWH 5VV FPVf DW WKH LQWHUIDFH WUDSV LQ WKH VKDGHG DUHD UHJLRQ LV SVV a FQVFSV1636 HQVHSV FQV16 HQV FSV36 HS FQVFSVA636 HQVHSV &QV16 HQV &SV36 HS 1LW[f [ G[ 1 ,7 f f ZKHUH FQV DQG FSV DUH WKH HOHFWURQ DQG KROH FDSWXUH UDWH FRHIILFLHQWV FPVHFf HQV DQG HSV DUH HPLVVLRQ UDWH FRHIILFLHQWV RI WKH WUDSSHG HOHFWURQ DQG KROH VHFf 1V DQG 3V DUH WKH FRQFHQWUDWLRQV RI HOHFWURQ DQG KROHV DW WKH LQWHUIDFH FPf DQG 1,7 LV WKH DUHDO GHQVLW\ RI LQWHUIDFH WUDS FPf $W DQG QHDU WKHUPDO HTXLOLEULXP LH QHJOHFW HOHFWULF ILHOG RU KRW FDUULHU HIIHFWf HQV FDQ EH UHODWHG WR FQV E\ QL H[S 6LPLODUO\ IRU HSV DQG F'6 SVf nSV Qs H[S (W (M N%7 (M f§ (W f QL f nSV N%7 f 3L f ,Q (TQVf DQG f Qc LV WKH LQWULQVLF FDUULHU FRQFHQWUDWLRQ RI VLOLFRQ (M LV WKH LQWULQVLF HQHUJ\ OHYHO N% LV WKH %ROW]PDQ FRQVWDQW DQG 7 LV WKH DPELHQW WHPSHUDWXUH

PAGE 43

6L LQWHUIDFH WUDSV LQ WKLV WKLQ LQWHUIDFLDO OD\HU )LJXUH &URVVVHFWLRQDO YLHZ RI D WKLQ LQWHUIDFLDO OD\HU DW WKH 6L6L LQWHUIDFH

PAGE 44

6XEVWLWXWLQJ (TQVf DQG f LQWR (Tf DQG XVLQJ QS QM 5VV FDQ EH VLPSOLILHG DV 1636 ULLf 5VV f§ f§ 1LW f FSV 1V Q[f FQV 3V S[f 7KH VXUIDFH UHFRPELQDWLRQ FXUUHQW ,%6 LV WKHQ WKH DUHDO LQWHJUDWLRQ RI WKH VXUIDFH UHFRPELQDWLRQ UDWH 5VV DORQJ WKH \ DQG ]GLUHFWLRQ OEV 5VV G\ G] f (TXDWLRQV f DQG f VKRZ WKDW WKHUH DUH WKUHH IDFWRUV WKDW ZLOO DIIHFW WKH FKDUDFWHULVWLFV RI WKH VXUIDFH UHFRPELQDWLRQ FXUUHQWV 7KH\ DUH f WKH VSDWLDO LQWHUIDFH WUDS GHQVLW\ 1,7\]f f WKH VSDWLDO VXUIDFH HOHFWURQ DQG KROH FDUULHU GHQVLWLHV 1V\]f DQG 3V\]f DQG f WKH HOHFWURQ DQG KROH FDSWXUH DQG HPLVVLRQ FRHIILFLHQWV FQV FSV HQV DQG HSV DQG WKH HQHUJ\ OHYHO RI WKH LQWHUIDFH WUDS LQ WKH VLOLFRQ HQHUJ\ JDS (M 'XULQJ KRW FDUULHU VWUHVV DQ LQFUHDVH LQ WKH GHQVLW\ RI WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDSV ZLOO LQFUHDVH WKH PDJQLWXGH RI WKH VXUIDFH UHFRPELQDWLRQ FXUUHQWV 7KLV LV WKH EDVLV RI WKH '&,9 ZKLFK HPSOR\V WKH VXUIDFH UHFRPELQDWLRQ FRPSRQHQW WR PRQLWRU KRWFDUULHU LQGXFHG GHJUDGDWLRQ LQ 0267V 7KH FDSWXUH DQG HPLVVLRQV FRHIILFLHQWV DUH DVVXPHG QRW WR EH DIIHFWHG E\ WKH KRWFDUULHU VWUHVV 6HYHUDO VDOLHQW FKDUDFWHULVWLF SURSHUWLHV RI WKH VXUIDFH UHFRPELQDWLRQ UDWH 5VV DUH EHVW XQGHUVWRRG E\ FRQVLGHULQJ WKH ,%F FRPSRQHQWV VKRZQ LQ )LJXUHV DQG ,Q WKH IROORZLQJ ZH ZLOO XVH WKH ,%BFO FRPSRQHQW PHDVXUHG XVLQJ WKH %RWWRP(PLWWHU '&,9 FRQILJXUDWLRQ DV DQ H[DPSOH LQ RXU GLVFXVVLRQ &RQVLGHU WKH WUDQVLWLRQ HQHUJ\

PAGE 45

EDQG GLDJUDP RI D Q0267 VKRZQ LQ )LJXUH ELDVHG LQ WKH ERWWRPHPLWWHU '&,9 FRQILJXUDWLRQ WKH JDWH YROWDJH PRGXODWHV WKH DPRXQW RI VXUIDFH HQHUJ\ EDQG EHQGLQJ ?MVf DW WKH 6L6L LQWHUIDFH DQG WKH IRUZDUGELDVHG ERWWRP HPLWWHU LQMHFWV PLQRULW\ FDUULHUV HOHFWURQf LQWR WKH SEDVH ZHOO 7KH UDWH RI HOHFWURQKROH UHFRPELQDWLRQ DW WKH LQWHUIDFH WUDSV 5VV ZLOO GHSHQG RQ WKH VXUIDFH HOHFWURQ DQG KROH FRQFHQWUDWLRQV 1V DQG 3V $V LQGLFDWHG LQ )LJXUH WKH\ DUH JLYHQ E\ 1 V DQG 1 QL H[S H[S H[S H[S T)1 )Sf N%7 N%7 IT\V Y) )1 )3f N%7 TAV 9) 9%( f N%7 f f f TIV 3V 3( H[S f f N%7 ,Q (TQV f DQG f 1( DQG 3( DUH WKH HTXLOLEULXP HOHFWURQ DQG KROH FRQFHQWUDWLRQV /RZ LQMHFWLRQ OHYHO LV DOVR DVVXPHG LH 1 DQG 3 3H 3H QLH[ST?_)N%7f )LJXUH LOOXVWUDWHV WKH UHODWLRQ EHWZHHQ WKH VXUIDFH UHFRPELQDWLRQ UDWH DQG VXUIDFH SRWHQWLDO DW WKH 6L6L LQWHUIDFH E\ SORWWLQJ 5VV YHUVXV ?_V IRU IRXU GLIIHUHQW IRUZDUGELDV YROWDJHV 7KH VKDSH RI WKH 5VV FXUYHV LV WKH VDPH EXW WKH PDJQLWXGH RI 5VV GHSHQGV H[SRQHQWLDOO\ RQ WKH H[FHVV VXUIDFH FDUULHU HOHFWURQf GHQVLW\ 7KH VXUIDFH

PAGE 46

6SDFHFKDUJH OD\HU Q 3RO\ 6L 6L S6L Q 6L *DWH 2[LGH %DVH (PLWWHU )LJXUH &URVVVHFWLRQDO YLHZ RI WKH WUDQVLWLRQ HQHUJ\ EDQG GLDJUDP

PAGE 47

gV9f )LJXUH 5HODWLRQ RI 5VV DQG ?MUV ZLWK 9%( DV SDUDPHWHUV $VVXPHG PLGHQHUJ\ OHYHO LQWHUIDFH WUDS LQ 6L HQHUJ\ JDS DQG FQV FSV 1$$ FP 7 .

PAGE 48

SRWHQWLDO DW ZKLFK WKH UHFRPELQDWLRQ UDWH SHDNV LV QHJDWLYHO\ GLVSODFHG DV WKH VXUIDFH PLQRULW\ FDUULHU HOHFWURQVf GHQVLW\ LQFUHDVHV GXH WR WKH LQMHFWLRQ RI PLQRULW\ FDUULHUV IURP WKH IRUZDUGELDVHG HPLWWHU 7KH VXUIDFH SRWHQWLDO 9_6PD[ FRUUHVSRQGV WR SHDN VXUIDFH UHFRPELQDWLRQ UDWH LV >@ 9AVPD[ YI YEHA f NJ7Tf ORJ FSVFQVf f $W WKLV VXUIDFH SRWHQWLDO WKH VXUIDFH LV LQWULQVLF LH WKH HOHFWURQ DQG KROH FRQFHQWUDWLRQV DUH WKH VDPH DQG HTXDO WR A%( 1V 3V QL H[S f f N%7 $VVXPLQJ (S (c F FQV FSV DQG 3V 1V Qc &2A,7QL AA%( 5VV f H[S f f N%7 7KLV VLPSOLILFDWLRQ LPSOLHV WKDW XQGHU PD[LPXP UHFRPELQDWLRQ FRQGLWLRQ ,%EHF A H[ST9%(N%7f 1H[W FRQVLGHU WKH FDVH IRU 3V 1V Qc 3VV FR ALW 1V f TOUV9)9%( &R1SJULSH[SO f f N%7 +HUH ,%%(BFO a H[ST9%(ON%7f 5HODWLRQVKLS EHWZHHQ ?_V DQG 9*% 8QGHU 1RQ(TXLOLEULXP &RQGLWLRQ 6R IDU ZH KDYH RQO\ FRQVLGHUHG WKH UHODWLRQVKLS RI 5VV DV D IXQFWLRQ RI 9_V ,W LV QHFHVVDU\ WR EH DEOH WR UHODWH WKH DSSOLHG JDWH YROWDJH 9*%f GXULQJ '&,9 PHDVXUHPHQW WR WKH VXUIDFH EDQGEHQGLQJ SRWHQWLDO 9_LV 7KLV FDQ EH REWDLQHG E\ PRGLI\LQJ WKH ERXQGDU\ FRQGLWLRQV ZKHQ VROYLQJ WKH 3RLVVRQ (TXDWLRQ (TQ f IRU 9M6

PAGE 49

eVLG(G[f eVL G9G[f T 313E 1Ef f ZKHUH e6L LV WKH VLOLFRQ GLHOHFWULF FRQVWDQW ( LV WKH HOHFWULF ILHOG VWUHQJWK LQWR WKH EXON 6L 3 LV WKH KROH FDUULHU FRQFHQWUDWLRQ 1 LV WKH HOHFWURQ FDUULHU FRQFHQWUDWLRQ 3% LV WKH EXON KROH FRQFHQWUDWLRQ DQG 1% LV WKH EXON HOHFWURQ FRQFHQWUDWLRQ 7KH ERXQGDU\ FRQGLWLRQV DW WKH 6L6L LQWHUIDFH DQG LQ WKH VLOLFRQ EXON DUH JLYHQ DV 9[ f 9V f 9; Rrf f ([ f (V f ( [ rRf f 3[ f SV f TAV 3%H[S f f N%7 3 ; FRf 3% f TY) +L H[S f f N%7 1[ f 1V f TAV 1%H[S f f N%7 1 ;f§rrf 1E f TY) TY%( UAH[S H[S f f N%7 N%7 8VLQJ WKH DERYH ERXQGDU\ FRQGLWLRQV DQG VROYLQJ (Tf E\ TXDGUDWXUH G9G[ G9G[fGG9fG9G[f fGG9fG9G[f f G(G9f f

PAGE 50

7KH VXUIDFH HOHFWULF ILHOG (V FDQ EH REWDLQHG DV ([ f (YVf f ZKHUH 8V 8) DQG 8%( DUH WKH QRUPDOL]HG SRWHQWLDO ZLWK UHVSHFW WR N%7T 8V T?VN%7 8) T9SN%7 DQG 8%( T9%(N%7f 7KH UHODWLRQVKLS EHWZHHQ WKH DSSOLHG JDWH YROWDJH 9*% DQG WKH VXUIDFH EDQGn EHQGLQJ SRWHQWLDO ?MV LV REWDLQHG E\ XVLQJ WKH 026& YROWDJH HTXDWLRQ f 9*% 9)% 9V B A2; e6L(V2; ZKHUH 9)% LV WKH IODWEDQG YROWDJH 9R[ LV WKH R[LGH YROWDJH DQG &R[ LV WKH R[LGH FDSDFLWDQFH SHU XQLW DUHD 7KXV IRU D JLYHQ 9*% (Tf FDQ EH VROYHG LWHUDWLYHO\ IRU ?_V 7KUHH H[DFW DV\PSWRWLF VROXWLRQV IRU (Tf FDQ EH GHULYHG E\ ZHLJKWLQJ WKH UHODWLYH PDJQLWXGH RI WKH FDUULHU FRQFHQWUDWLRQV DW WKH 6L6L LQWHUIDFH XQGHU VXUIDFH DFFXPXODWLRQ GHSOHWLRQ DQG LQYHUVLRQ FRQGLWLRQV >@ 8QGHU VXUIDFH DFFXPXODWLRQ FRQGLWLRQ ZKHUH 8V 9*% 9IE 9nVf 9V N%7TfORJH $ f$ f N%7Tf9$$ 8QGHU VXUIDFH GHSOHWLRQ FRQGLWLRQ ZKHUH 8V 8) 8%( AJE AIE N%7Tf$' 9V a A*% 9S%9$$ f 8QGHU VXUIDFH LQYHUVLRQ FRQGLWLRQ ZKHUH 8V 8) 8%(

PAGE 51

9V 9I9EH N%7 ORJ A*% A)% 9Vf N%7Tf9$$$, f ,Q (TQV ff WKH LWHUDWLYH FRUUHFWLRQ WHUPV $D $G DQG $Mf DUH $D 86>H[S 8Vf86@ f H[S 8I8EHf f $G H[S 8Vf >H[S XVf86@ }H[S 8I8EHf f $M 86f H[S 8Vf >H[S 86f 86@ f 8VLQJ (TQVff WKH UHODWLRQVKLS EHWZHHQ ?_V DQG 9*% LV FRPSXWHG IRU [; QP 1$$ FP 9)% 9 DQG 7 7KH UHVXOW LV VKRZQ LQ )LJXUH ZLWK 9%( DV SDUDPHWHU ,Q WKH DFFXPXODWLRQ DQG GHSOHWLRQ UDQJHV WKH UHVXOWV FRPSXWHG IRU QRQ]HUR 9%(fV DUH VLPLODU WR WKH FDVH ZLWK 9%( 9 7KLV LQGLFDWHV WKDW XQGHU ORZWRPRGHUDWH IRUZDUGELDVHV WKH H[FHVV VXUIDFH HOHFWURQ FRQFHQWUDWLRQ GXH WR HOHFWURQ LQMHFWLRQ IURP WKH IRUZDUGELDVHG HPLWWHU MXQFWLRQ LV QRW DSSUHFLDEOH WR DIIHFW WKH ILHOG DQG SRWHQWLDO GLVWULEXWLRQ LQ WKH VXUIDFH VSDFHFKDUJHOD\HU +RZHYHU LQ WKH VWURQJ LQYHUVLRQ UDQJH H[FHVV VXUIDFH HOHFWURQ FRQFHQWUDWLRQ UHGXFHV WKH DPRXQW RI VXUIDFH EDQG EHQGLQJ E\ 9%( 9V LQY f 9I 9EH f 7KH UHVXOWV VKRZQ LQ )LJXUHV DQG DUH FRPELQHG DQG UHSORWWHG LQ )LJXUH ZKLFK LOOXVWUDWHV WKH JDWHYROWDJH GHSHQGHQFH RI VXUIDFH UHFRPELQDWLRQ UDWH 5VV IRU IRXU IRUZDUGELDV FRQGLWLRQV 9%( DQG 9f 7KLV WKHRUHWLFDO FDOFXODWLRQV LQGLFDWHV WKDW EDVH FXUUHQW PHDVXUHG GXULQJ '&,9 ZLOO H[KLELW D PD[LPXP ZKHQ LWV VXUIDFH UHFRPELQDWLRQ FRPSRQHQWV GRPLQDWH )RU D VLQJOH HQHUJ\OHYHO LQWHUIDFH WUDS

PAGE 52

9*%9f )LJXUH 5HODWLRQ EHWZHHQ VXUIDFH EDQGEHQGLQJ SRWHQWLDO \Vf DQG DSSOLHG JDWH YROWDJH 9*%f XQGHU HTXLOLEULXP 9%( 9f DQG QRQHTXLOLEULXP 9%( DQG 9f FRQGLWLRQV 3DUDPHWHUV XVHG [R[ QP 1$$ FP 9UD 9 DQG 7 .

PAGE 53

9*%9f )LJXUH 5HODWLRQ EHWZHHQ VXUIDFH UHFRPELQDWLRQ UDWH 5VVf DQG DSSOLHG JDWH YROWDJH 9*%f XQGHU HTXLOLEULXP 9%( 9f DQG QRQHTXLOLEULXP 9%( DQG 9f FRQGLWLRQV 3DUDPHWHUV XVHG [R[ QP 1$$ FP 9)% 9 1Q FPn DQG 7 .

PAGE 54

WKH ,% SHDN RFFXUV ZKHQ WKH VXUIDFH HOHFWURQ DQG KROH GHQVLWLHV DUH FRPSDUDEOH WR HDFK RWKHU 6XPPDU\ $ GF FXUUHQWYROWDJH '&,9f WHFKQLTXH LV LQWURGXFHG LQ WKLV FKDSWHU WR LQYHVWLJDWH KRWFDUULHULQGXFHG GHJUDGDWLRQ LQ 026 WUDQVLVWRUV 7KLV WHFKQLTXH XVHV WKH EDVH FXUUHQW ,% LQ D JDWHFRQWUROOHG SDUDVLWLF %-7 RI D 0267 WR PRQLWRU WKH VWUHVVJHQHUDWHG R[LGH DQG LQWHUIDFH WUDSV ,W KDV VHYHUDO XQLTXH IHDWXUHV f LW LV D SXUH GF PHDVXUHPHQW ZKLFK JUHDWO\ VLPSOLILHV LQVWUXPHQWDWLRQ VHWXS f LW LV GLUHFWO\ DSSOLFDEOH WR VXEPLFURQ DUHD GHYLFHV LQ FRQYHQWLRQDO SURGXFWLRQ &026EDVHG WUDQVLVWRU VWUXFWXUH DQG FDQ PRQLWRU WKH WUDQVLVWRUfV GHJUDGDWLRQ NLQHWLFV XQGHU KRWFDUULHU VWUHVV f WKH PDJQLWXGH RI ,% LV GLUHFWO\ SURSRUWLRQDO WR WKH LQWHUIDFH WUDS GHQVLW\ 1,7 RU 4,;T DQG WKXV QRW DIIHFWHG E\ LQKRPJHQLW\ RI 1,7 f WKH VXUIDFH UHFRPELQDWLRQ UDWH 5VV DQG PLQRULW\ LQMHFWLRQ OHYHO FDQ EH LQGHSHQGHQWO\ FRQWUROOHG E\ WKH DSSOLHG JDWH YROWDJH DQG IRUZDUG ELDV YROWDJHV UHVSHFWLYHO\ DQG f LW KDV KLJK VHQVLWLYLW\ ZKLFK JLYHV D PLQLPXP PHDVXUDEOH 1,7 RI WUDSVFP )LJXUH f 7KH '&,9 WHFKQLTXH KDV EHHQ UHFHQWO\ SDFNDJHG LQWR D WXUQNH\ GLVNHWWH 6$0'&,9 >@f IRU WKH +3 '& 3DUDPHWHU $QDO\]HUf D W\SLFDO FKDUDFWHUL]DWLRQ HTXLSPHQW XVHG DPRQJ VHPLFRQGXFWRU FRPSDQLHV 7KLV GLVNHWWH WRJHWKHU ZLWK D XVHU JXLGH DQG WXWRULDO GRFXPHQWDWLRQ KDV EHHQ GLVWULEXWHG WR PHPEHUV RI WKH 6HPLFRQGXFWRU 5HVHDUFK &RUSRUDWLRQ DV SDUW RI D WHFKQRORJ\ WUDQVIHU HIIRUW

PAGE 55

L 2 &' 2 L L L L O L L L L L L L L L L L L Wf§Lf§Lf§Lf§_f§Lf§Lf§Lf§Lf§_f§Lf§Lf§Lf§Lf§_f§Lf§Lf§Lf§U 7RS(PLWWHU %-7 f $,E T$6QfH[ST9%(N7f S$ f 6 FPVHF 17 FUUI L L L L O L L L L L L L L L L L LB 9*% 9f )LJXUH 6HQVLWLYLW\ WHVW RI WKH WRSHPLWWHU ,% PHDVXUHPHQW RI LQWHUIDFH WUDSV RI DQ XQVWUHVVHG %L026 ZLWK Q0267 DQG QSQ %-7 DW 9(% 9 DQG ;T; QP :/ _LP LP

PAGE 56

&+$37(5 6(3$5$7,21 2) ,17(5)$&( $1' 2;,'( 75$36 86,1* 7+( '&,9 0(7+2' ,QWURGXFWLRQ (OHFWULFDOO\ DFWLYH R[LGH 427f DQG LQWHUIDFH WUDSV 4,7f DUH NQRZQ WR DIIHFW WKH HOHFWULFDO FKDUDFWHULVWLFV RI PHWDOR[LGHVHPLFRQGXFWRU FDSDFLWRU 026&f DQG PHWDO R[LGHVHPLFRQGXFWRU WUDQVLVWRUV 0267f $UHDOO\ XQLIRUP DQG FKDUJHG R[LGH WUDSV LQ 026& DQG 0267 ZLOO FDXVH WKH &J9* DQG ,TVn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f DQG GF ,'9* WUDQVIHU FKDUDFWHULVWLFVf KDYH EHHQ URXWLQHO\ XVHG WR VWXG\ WKH JHQHUDWLRQ DQG FKDUJLQJ RI R[LGH DQG LQWHUIDFH WUDSV GXULQJ KRW FDUULHU VWUHVV )RU WKH WZRWHUPLQDO KLJK IUHTXHQF\ &J9* PHWKRG WKH VWUHVVFUHDWHG R[LGH WUDS GHQVLW\ LV PRQLWRUHG E\ WKH FKDQJH LQ IODWEDQG YROWDJH ZKLOH WKH LQWHUIDFH WUDS GHQVLW\ LV FDOFXODWHG IURP H[SHULPHQWDO GDWD XVLQJ WKH 7HUPDQ PHWKRG 7KLV PHWKRG UHTXLUHV D ODUJH DUHD 026&

PAGE 57

WHVW VWUXFWXUH IRU DFFXUDWH PHDVXUHPHQW RI WKH VPDOOVLJQDO FDSDFLWDQFH DQG WKXV LV QRW VXLWDEOH IRU WHVW VWUXFWXUH ZLWK YHU\ VPDOO WUDQVLVWRUV )RU WKH GF ,'9* PHWKRG WKH H[WUDFWHG WKUHVKROG YROWDJH VKLIW FDQ EH XVHG WR PRQLWRU WKH GHQVLW\ RI WKH R[LGH WUDSV SURYLGHG WKDW WKH R[LGH WUDS GHQVLW\ LV DUHDOO\ XQLIRUP DQG WKH LQWHUIDFH WUDS GHQVLW\ LV VXIILFLHQWO\ ORZ 7KH LQWHUIDFH WUDS GHQVLW\ FDQ EH TXDOLWDWLYHO\ REWDLQHG IURP WKH VXEWKUHVKROG VORSH FKDQJH +RZHYHU LI ERWK R[LGH DQG LQWHUIDFH WUDSV DUH JHQHUDWHG GXULQJ WKH KRW FDUULHU VWUHVV H[SHULPHQW WKHQ LW LV QRW HDV\ WR VHSDUDWH WKH UHODWLYH FRQWULEXWLRQ RI 427 DQG 4,7 IURP D VLQJOH ,'9* PHDVXUHPHQW EHFDXVH WKH WRWDO JDWH YROWDJH VKLIW LV GXH WR WKH EXLOGXS RI ERWK 427 DQG 4,7 $9J $9JBRW $9JBL7 f $4USUS $4LWf ZKHUH $9*27 LV D FRPSRQHQW RI WKH WRWDO JDWH YROWDJH VKLIW $9* FDXVHG E\ WKH FKDUJHG R[LGH WUDS %\ WKH VDPH DUJXPHQW WKH LQWHUIDFH WUDS FRQWULEXWHV D $9*,7 FRPSRQHQW ,Q DGGLWLRQ ,' DOVR GHSHQGV RQ PRELOLW\ ZKLFK LV FKDQJHG E\ 4,7 DQG 427 ,Q WKLV FKDSWHU ZH ZLOO VKRZ WKDW WKH '&,9 PHWKRG FDQ EH XVHG WR VHSDUDWH WKH $4RW DQG $4,7 >@ 7KLV QRYHO PHWKRG FRQWDLQV WZR IHDWXUHV f WKH LQFUHDVH LQ WKH PDJQLWXGH RI EDVH FXUUHQW LV XVHG WR PRQLWRU WKH VWUHVVJHQHUDWHG 4,7 f WKH ODWHUDO YROWDJH VKLIW RI WKH FROOHFWRU FXUUHQW YHUVXV JDWH YROWDJH LV XVHG WR GHWHFW WKH 427 ([SHULPHQW 7R GHPRQVWUDWH WKH VHSDUDWLRQ RI 4,7 DQG 47 XVLQJ WKH '&,9 PHWKRG D Q0267 IDEULFDWHG ZLWK D MXQFWLRQLVRODWHG SZHOO LV VXEMHFWHG WR 6XEVWUDWH +RW (OHFWURQ

PAGE 58

LQMHFWLRQ 6+(Lf WR XQLIRUPO\ JHQHUDWH 4,7 DQG 4TW 7KH WUDQVLVWRU KDV D GUDZQ FKDQQHO OHQJWK RI _LP DQG D ZLGWK RI _LP 7KH JDWH R[LGH WKHUPDOO\ JURZQ DW r& LQ D GU\ R[LGDWLRQ DPELHQW LV $ WKLFN $ FURVVVHFWLRQDO YLHZ RI WKH Q0267 LV VKRZQ LQ )LJXUH 7KH VWUHVVJHQHUDWHG 4,7 DQG 427 DUH PRQLWRUHG E\ WKH WRS DQG ERWWRPHPLWWHU '&,9 DQG WKH FRQYHQWLRQDO ,'9* PHDVXUHPHQWV 7KH WRS DQG ERWWRPHPLWWHU PHDVXUHPHQW FRQILJXUDWLRQV DUH VKRZQ LQ )LJXUH )LJXUH VKRZV WKH HQHUJ\ EDQG EHQGLQJ RI WKH JDWH R[LGH DQG WKH 6L VXEVWUDWH LQ WKH Q0267 GXULQJ 6+( LQMHFWLRQ $ SRVLWLYH YROWDJH 9* 9f LV DSSOLHG EHWZHHQ WKH JDWH DQG WKH VRXUFH DQG GUDLQ MXQFWLRQV ZKLFK DUH FRQQHFWHG WRJHWKHU WR LQYHUW WKH FKDQQHO XQGHUQHDWK WKH JDWH R[LGH 7KLV LQYHUWHG QFKDQQHO WRJHWKHU ZLWK WKH Q VRXUFH DQG GUDLQ MXQFWLRQV VHUYHV DV WKH FROOHFWRU IRU WKH YHUWLFDO QSQFKDQQHO ELSRODU WUDQVLVWRU 7KH SZHOO LV WKH EDVH DQG WKH QQ HSLWD[LDO VXEVWUDWH VHUYHV DV WKH HOHFWURQ HPLWWHU 7KH HPLWWHUEDVH MXQFWLRQ LV IRUZDUGELDVHG WR LQMHFW HOHFWURQV LQWR WKH SZHOO EDVH 7KHVH LQMHFWHG HOHFWURQV GLIIXVH XSZDUGV WKURXJK WKH SZHOO DQG DUH DFFHOHUDWHG WRZDUGV WKH 6L6L LQWHUIDFH E\ WKH HOHFWULF ILHOG LQ WKH VXUIDFH VSDFHFKDUJH OD\HU RI WKH UHYHUVHELDVHG FROOHFWRUEDVH QFKDQQHOSZHOO MXQFWLRQ 7KH UHYHUVH ELDV 9&% SURYLGHV WKH DFFHOHUDWLQJ ILHOG IRU WKHVH HOHFWURQV :KHQ WKH FKDQQHO LV VWURQJO\ LQYHUWHG 9&% ZLOO EH WKH SRWHQWLDO GLIIHUHQFH EHWZHHQ WKH KROH TXDVL)HUPL SRWHQWLDO 9S )3T LQ WKH SZHOO DQG WKH HOHFWURQ TXDVL)HUPL SRWHQWLDO 91 )1Tf DW WKH 6L6L LQWHUIDFH 7KH DFFHOHUDWLQJ HOHFWURQV PD\ DFTXLUH VXIILFLHQW HQHUJ\ LQ WKH 6L VSDFHFKDUJH OD\HU WR VXUPRXQW WKH 6L6L EDUULHU H9f +HQFH WKH PLQLPXP 9&% WKDW PXVW EH DSSOLHG WR DFFHOHUDWH WKH LQMHFWHG HOHFWURQV LQWR WKH JDWH R[LGH LV DERXW

PAGE 59

6RXUFH *DWH 'UDLQ %DVH Q 0 ‘‘ Q SEDVH FRQWDFW SEDVH ZHOO QHSLWD[LDO OD\HU Q VXEVWUDWH 6XEVWUDWH )LJXUH &URVVVHFWLRQDO YLHZ RI WKH %L026 %-70267 WUDQVLVWRU VWUXFWXUH XVHG

PAGE 60

7RS(PLWWHU &RQILJXUDWLRQ %RWWRP(PLWWHU &RQILJXUDWLRQ &W 29 RY 6L2 Q&RIOHFWRU (% Â’ 4_7 Â’ Er27 $\LIDL L UQUWL % RY S%DVF &RQWDF S%DVH QHSLWD[LDO OD\HU Q(PLWWHU ,H 9 )LJXUH &URVVVHFWLRQDO YLHZV RI WKH WRSHPLWWHU DQG ERWWRPHPLWWHU '&,9 PHDVXUHPHQW FRQILJXUDWLRQV 7KH ELDV XVHG LQ WKH H[SHULPHQW DUH DOVR GHSLFWHG LQ WKH ILJXUH

PAGE 61

(OHFWURQV )LJXUH 7UDQVLWLRQ HQHUJ\ EDQG GLDJUDP RI %L026 VXEVWUDWH KRW HOHFWURQ LQMHFWLRQ

PAGE 62

9 ,Q WKH 6+(L H[SHULPHQW D UHYHUVH ELDV RI 9 LV FKRVHQ IRU 9&% 6RPH RI WKH HOHFWURQV WKDW DUH LQMHFWHG LQWR WKH 6L FDQ EH FDSWXUHG E\ WKH QHXWUDO R[\JHQ YDFDQF\ FHQWHUV >@ JLYLQJ 9* Hf 9 f DQG UHVXOWV LQ D QHW EXLOWXS RI QHJDWLYH 47 %HFDXVH RI WKH KLJK NLQHWLF HQHUJ\ aH9 IURP 9VE 9GE 9f RI WKH DFFHOHUDWLQJ HOHFWURQV LQ WKH VXUIDFH VSDFHFKDUJHOD\HU WKH\ FDQ DOVR EUHDN WKH ZHDN RU VWUDLQHG LQWHUIDFLDO ERQGV 6L6L 6L2 6L+ 6L2+f YLD GLUHFW LPSDFW WR FUHDWH QHZ LQWHUIDFH WUDSV >@ VLQFH WKH ERQG HQHUJ\ RI RI WKHVH LQWHUIDFLDO ERQG LV RI WKH RUGHU RI H9 'XULQJ 6+(L WKH JDWH R[LGH FXUUHQW LV NHSW FRQVWDQW ZLWKLQ sb RI WKH LQLWLDO SUHVHW JDWH FXUUHQW E\ GLJLWDO IHHGEDFN FRQWURO RI WKH HPLWWHUEDVH IRUZDUG ELDV 7KH R[LGH HOHFWULF ILHOG (R[ LV JLYHQ E\ 9* 9Vf[R[ 7KH JDWH HOHFWURQ IOXHQFH )* RU 1LQM LV FDOFXODWHG DV IROORZV 1LQM -r -*Tf GW f -*TfrW VWUHVV f ZKHUH -* LV WKH FRQVWDQW JDWH FXUUHQW GHQVLW\ T LV WKH HOHFWURQLF FKDUJH O[O2n &f DQG WVWUHVV LV WKH WRWDO VWUHVV WLPH LQ VHFRQGV 2QH NH\ IHDWXUH RI XVLQJ WKH 6+(L WHFKQLTXH LV WKDW WKH WKUHH EDVLF VWUHVV SDUDPHWHUV WKH HOHFWURQ NLQHWLF HQHUJLHV LQ WKH 6L DQG DW WKH 6L6L LQWHUIDFH DQG WKH LQMHFWLRQ FXUUHQW FDQ EH LQGHSHQGHQWO\ FRQWUROOHG > @ 7KH HOHFWURQ NLQHWLF HQHUJ\ LQ WKH 6L LV FRQWUROOHG E\ WKH YROWDJH DSSOLHG EHWZHHQ WKH JDWH WHUPLQDO DQG WKH QVRXUFHQGUDLQ WHUPLQDOV SURYLGHG WKH JDWH YROWDJH LV VXIILFLHQW WR LQYHUW WKH VLOLFRQ

PAGE 63

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f WKH LQFUHDVH LQ WKH PDJQLWXGH RI WKH SRVWVWUHVVHG ,% SHDN LV DWWULEXWHG WR DQ LQFUHDVH LQ WKH GHQVLW\ RI WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDS 7KH 9*%LW FRPSRQHQW LQ (TQ,f FDQ WKHQ EH GHWHUPLQHG XVLQJ WKLV VWUHVVLQGXFHG ,% $,% ,%SRVWVWUHVVf ,%SUHVWUHVVf XVLQJ WKH IROORZLQJ UHODWLRQV >@ DLE T$A$6Af H[S T9%(N%7f f $6 7 f 7RWK f $1LW f $4LW f§&T$9JELW f T$1,7 f T$',7m $(LW f $6 LV WKH LQFUHDVH LQ VXUIDFH UHFRPELQDWLRQ YHORFLW\ GXH WR WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDSV $* LV WKH JDWH DUHD RI WKH Q0267 Qc LV WKH LQWULQVLF FDUULHU FRQFHQWUDWLRQ WK LV FDUULHU WKHUPDO YHORFLW\ ,Q HTQV f DQG f LW LV DVVXPHG WKDW WKH GHQVLW\RIVWDWH '7f DQG WKH FDUULHU FDSWXUH FURVV VHFWLRQV DQ DS Rf RI WKH

PAGE 64

)LJXUH (IIHFW RI DUHDOXQLIRUP 6+(L VWUHVV RQ WKH QSQ %-7 DQG Q0267 FKDUDFWHULVWLFV DV D IXQFWLRQ RI 9*% $OO VWUHVVHG DW [O HOHFWURQ FP IOXHQFH H[FHSW ERWWRP( ,% LQ Df 0HDVXUHG DW 9(% 9 9&% 9 DQG Df %DVH FXUUHQW DQG Ef FROOHFWRU FXUUHQW LQ WKH WRSHPLWWHU DQG ERWWRPHPLWWHU FRQILJXUDWLRQV Ff 6XEWKUHVKROG ,G9JE RI WKH Q0267

PAGE 65

LQWHUIDFH WUDSV LQ WKH 6L HQHUJ\ JDS DUH LQGHSHQGHQW RI HQHUJ\ ,Q )LJXUH Df WKH ,JSHDN IRU WKH WRS(PLWWHU FXUYH VWUHVVHG DW DQ HOHFWURQ IOXHQFH RI [O2OFPn RFFXUV DW D VXUIDFH EDQGEHQGLQJ RI 9V 9) 9 DQG JLYHV $6 FPV 7KH ,%SHDN IRU WKH ERWWRP(PLWWHU FXUYH VWUHVVHG DW DQ HOHFWURQ IOXHQFH RI O[OFPn RFFXUULQJ DW ??LV 9) 9 JLYHV $6 FPVHF ,Q WKH DERYH H[DPSOH ?MV IrU WKH WRS( DQG ERWWRP( FXUYHV DUH GHWHUPLQHG XVLQJ WKH UHODWLRQV SUHVHQWHG LQ FKDSWHU ,Q )LJXUH Ef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f DUH LQGLFDWHG RQ WKH SUHVWUHVV ,*9*% FXUYH $V VHHQ IURP )LJXUH Df WKH 6+(L KDV JHQHUDWHG VXEVWDQWLDO LQWHUIDFH WUDSV LQGLFDWHG E\ WKH ODUJH LQFUHDVH LQ ,% %XW WKH SRVW

PAGE 66

VWUHVV ,F9*% FXUYH RQO\ H[KLELWV D ODWHUDO DQG DOPRVW SDUDOOHO VKLIW IURP WKH SUHVWUHVV FXUYH ZLWK QHDUO\ LGHQWLFDO KHLJKW IRU ERWK WKH WRS(PLWWHU DQG ERWWRP(PLWWHU PHDVXUHPHQW FRQILJXUDWLRQV 7KLV ODWHUDO SRVLWLYH JDWH YROWDJH VKLIW $9*% LV DWWULEXWHG WR WKH FKDUJLQJ RI WKH QHXWUDO R[\JHQ YDFDQF\ FHQWHUV DV VKRZQ LQ HTQf 7KXV WKH 9*%RW FRPSRQHQWV LQ (TQf FDQ EH GHWHUPLQHG IURP DT[W T f & Tf f $9T%B7 f &Tff> f@ f [O/FPn f 7KH HIIHFW RI KRWHOHFWURQLQGXFHG GHJUDGDWLRQ RQ WKH Q0267fV FKDUDFWHULVWLFV FDQ DOVR EH REVHUYHG IURP WKH FRQYHQWLRQDO ,G9JE PHDVXUHPHQW VKRZQ LQ )LJXUH Ff $IWHU LQMHFWLQJ WKH JDWH R[LGH ZLWK DQ HOHFWURQ IOXHQFH RI [OFPf WKH SRVW VWUHVVHG ,' FXUUHQW GHFUHDVHV ZLWK LQFUHDVLQJ 9*% DQG WKHUH LV FOHDUO\ D GLVWRUWLRQ LQ WKH VXEWKUHVKROG UHJLPH 7KH VXEWKUHVKROG VORSH FKDQJH $6 FDQ EH HPSOR\HG WR HVWLPDWH WKH GHQVLW\ RI LQWHUIDFH WUDSV XVLQJ WKH IROORZLQJ ZHOONQRZQ HTXDWLRQ >@ 'LW &Tf TN%7f $6 f FPBH9 f +RZHYHU WKH ,' GHJUDGDWLRQ LV FOHDUO\ GXH WR WKH VWUHVVJHQHUDWHG $4,7 DQG $47 ZKRVH SUHVHQFH LV XQDPELJRXVO\ FRQILUPHG E\ WKH ,%9*% DQG ,&9*% PHDVXUHPHQW 7KLV LV GHSLFWHG LQ )LJXUH Ff ZKHUH WKH DPRXQW RI WRWDO JDWH YROWDJH VKLIW $9*% H[WUDFWHG DW D FRQVWDQW GUDLQ FXUUHQW LV VKRZQ WR FRQVLVWV RI WZR FRPSRQHQWV $9*%B7 DQG $9JEBLW 7KH VHSDUDWLRQ RI WKH WRWDO JDWH YROWDJH VKLIW $9*% LQWR WKHVH WZR

PAGE 67

FRPSRQHQWV FDQQRW EH DFFRPSOLVKHG IURP RQH VLQJOH ,*9*% PHDVXUHPHQW DORQH XQOHVV DGGLWLRQDO SURSHUWLHV RI WKH LQWHUIDFH WUDSV DUH NQRZQ RU DVVXPHG +RZHYHU WKLV FDQ EH HDVLO\ VHSDUDWHG XVLQJ WKH '&,9 WHFKQLTXH 2Q WKH ,%9*% SORW VKRZQ LQ )LJXUH Df WKH SRVWVWUHVVHG ,% SHDN LQFUHDVHV E\ DERXW WLPHV ZKLFK WUDQVODWH WR DERXW FPn 1LW DIWHU [O HOHFWURQFPn 6+( VWUHVV 2Q WKH ,FA*% VKRZQ LQ WKH )LJXUH Ef WKH SRVWVWUHVVHG ,F FXUYH VKLIWHG E\ DERXW 9 ZKLFK WUDQVODWHV WR D QHJDWLYHO\ FKDUJHG 47 GHQVLW\ RI [ FPn $GGLWLRQDO H[DPSOHV DUH JLYHQ LQ )LJXUH DfFf IRU WKH ERWWRP(PLWWHU FRQILJXUDWLRQ ZKLFK XVH ,F WR PRQLWRU QHJDWLYH SRVLWLYH DQG WXUQDURXQG $47 LQGXFHG E\ VWUHVV )LJXUH Df LV LGHQWLFDO WR )LJXUH Ef VKRZLQJ SRVLWLYH $9*% FDXVHG E\ WKH QHJDWLYH R[LGH WUDS 427 )LJXUH Ef VKRZV QHJDWLYH $9*% FXUYH f IURP SRVLWLYH R[LGH WUDS 47 7KH JHQHUDWLRQ RI SRVLWLYH R[LGH WUDS 4W DQWLFLSDWHG IURP WKH HOHFWURQLPSDFW HPLVVLRQ RI HOHFWURQV WUDSSHG E\ WKH QHXWUDO R[\JHQ YDFDQF\ >@ IRU D 6+(L FRQGLWLRQ ZLWK 9*% 9 DQG 9'% 96% 9 9r Hr 9 Ha f )LJXUH Ff LOOXVWUDWHV WKH 427 WXUQDURXQG HIIHFW DV FRLQHG E\ @ $ VKRUW &+(L VWUHVV aOVHFf DW 9*% 9'% 9 ZLWK WKH VRXUFH IORDWLQJ LQMHFWV HOHFWURQ LQWR WKH R[LGH DORQJ WKH HQWLUH OHQJWK RI WKH VWURQJO\ LQYHUWHG QFKDQQHO EHFDXVH 9*% } 97+ 7KH HOHFWURQV DUH FDSWXUHG E\ WKH H[LVWLQJ R[LGH WUDS UHVXOWLQJ LQ 427 &XUYH f $IWHU DQ DGGLWLRQDO VHF VWUHVV WKHUH LV D EXLOWXS RI SRVLWLYH R[LGH WUDSV IURP WKH LPSDFW HPLVVLRQ SDWKZD\ GHVFULEHG E\ HTQf ZKLFK FRPSHQVDWHV WKH 4W DQG VKLIWV WKH FXUYH WR WKH ULJKW &XUYH f

PAGE 68

9JE 9f )LJXUH (IIHFW RI VWUHVV RQ WKH FROOHFWRU FXUUHQW PHDVXUHG DW 9(% ERWWRPHPLWWHUf 9 9&% 9 DQG Df [O FP 6+(L VWUHVV DW 9*% 9 9&% 9 DQG ,* ,Q$ Ef [O FP 6+(L VWUHVV DW 9*% 9 9&% 9 DQG ,* ,Q$ Ff &+(L VWUHVV DW 9*% 9'% 9 DQG IORDWLQJ 96% DQG ,' O$ IRU VHF FXUYH f DQG V FXUYH f

PAGE 69

&+$37(5 352),/,1* 2) ,17(5)$&( 75$36 %< 7+( '&,9 0(7+2' ,QWURGXFWLRQ 'XH WR WKH ORFDOL]HG QDWXUH RI FKDQQHOKRWFDUULHU LQGXFHG GHJUDGDWLRQ PXOWLn GLPHQVLRQDO WUDQVLVWRU VLPXODWRUV VXFK DV 3,6&(6 RU 0,1,026 DUH XVXDOO\ HPSOR\HG WR DQDO\]H WKH HIIHFWV RI FKDQQHOKRWFDUULHU LQGXFHG GHJUDGDWLRQ RQ WKH HOHFWULFDO SHUIRUPDQFH RI WKH WUDQVLVWRUV > @ +RZHYHU ODFN RI UHOLDEOH H[SHULPHQWDO GDWD KDV RIWHQ UHTXLUHG XVLQJ DVVXPHG VSDWLDO GLVWULEXWLRQ RI LQWHUIDFH DQG R[LGH LQ WUDQVLVWRU VLPXODWLRQV +HQFH WKHUH LV D QHHG IRU D UHOLDEOH FKDUDFWHUL]DWLRQ WHFKQLTXH WR SURILOH WKH VSDWLDO GLVWULEXWLRQ RI VWUHVVJHQHUDWHG LQWHUIDFH DQG R[LGH WUDSV ZKHUH WKH H[SHULPHQWDOO\GHWHUPLQHG LQWHUIDFH DQG R[LGH WUDS SURILOH FDQ EH IHG LQWR WKH VLPXODWRUV WR DQDO\]H WKH HOHFWULFDO FKDUDFWHULVWLFV RI D GHJUDGHG WUDQVLVWRU 7ZR H[SHULPHQWDO WHFKQLTXHV ZHUH UHFHQWO\ SURSRVHG WR FKDUDFWHUL]H WKH VSDWLDO GLVWULEXWLRQ RI VWUHVVJHQHUDWHG LQWHUIDFH WUDSV 4,7f QHDU WKH GUDLQ MXQFWLRQ RI D Q 0267 >@ 6SHFNEDFKHU HW DO >@ XVHG WKH GF VXUIDFH JHQHUDWLRQ FXUUHQW WHFKQLTXH WR SURILOH WKH VSDWLDO GLVWULEXWLRQ RI 4,7 ,Q WKLV WHFKQLTXH WKH ZHDN UHYHUVH ELDVHG GUDLQ MXQFWLRQ FXUUHQW ZDV PHDVXUHG LQ D JDWHG GLRGH FRQILJXUDWLRQ DQG WKH JDWH ELDV LV YDULHG IURP DFFXPXODWLRQ WR LQYHUVLRQ WR YDU\ WKH VXUIDFH VSDFHFKDUJH OD\HU QHDU WKH GUDLQ MXQFWLRQ 'XH WR WKH VPDOO UHYHUVH ELDV DSSOLHG RQ WKH GUDLQ WKH SURILOLQJ

PAGE 70

FDSDELOLW\ RI WKLV WHFKQLTXH LV FRQILQHG WR UHJLRQ YHU\ FORVH WR WKH GUDLQ MXQFWLRQ )XUWKHUPRUH WKH VHQVLWLYLW\ ZDV OLPLWHG E\ YHU\ ORZ VWUHVVJHQHUDWHG FXUUHQWV a I$f VXSHULPSRVHG RQ ODUJH XQVWUHVVHG OHDNDJH FXUUHQW $QFRQD HW DO >@ +HUHPDQV HW DO >@ DQG 0D HW DO >@ HPSOR\HG WKH FKDUJHSXPSLQJ FXUUHQW WHFKQLTXH >@ ZKLFK LV D YDULDWLRQ RI WKH ZHOONQRZQ SXOVH ILHOGHIIHFW PHWKRG >@ 7KH VSDWLDO GLVWULEXWLRQ RI 4,7 LV REWDLQHG E\ FRUUHODWLQJ WKH GUDLQ YROWDJH SXOVH GXULQJ WKH FKDUJH ILOOLQJ DFFXPXODWLRQf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r RXWRISKDVHf WR DYRLG HUURQHRXV LQWHUSUHWDWLRQ RI WKH 4,7 VSDWLDO GLVWULEXWLRQ ,Q WKLV FKDSWHU ZH ZLOO VKRZ WKDW WKH '&,9 PHWKRG FDQ EH H[WHQGHG WR SURILOH VSDWLDO GLVWULEXWLRQ RI VWUHVVJHQHUDWHG 4UU W@ 7KHRU\ 7KH '&,9 3URILOLQJ 7HFKQLTXH HPSOR\V WKH ,%B& FRPSRQHQWV WR GHWHUPLQH WKH ODWHUDO GLVWULEXWLRQ RI WKH LQWHUIDFH WUDSV 7KH PDJQLWXGH RI ,%& LV WKH DUHDO LQWHJUDWLRQ RI WKH VXUIDFH UHFRPELQDWLRQ UDWH 5VV PXOWLSOLHG E\ WKH HOHFWURQLF FKDUJH T $VVXPLQJ D FRQVWDQW FKDQQHO ZLGWK : ,%& FDQ EH H[SUHVVHG DV

PAGE 71

A%,& T : /f§\V 5VV G< \' f ,Q )LJXUH WKH GUDLQERG\ PHWDOOXUJLFDO QS MXQFWLRQ LV UHIHUHQFHG DV WKH RULJLQ RI WKH \GLUHFWLRQ 7KXV \' LV WKH GUDLQ VXUIDFHFKDUJH OD\HU WKLFNQHVV \V LV WKH VRXUFH VXUIDFH VSDFHFKDUJH OD\HU WKLFNQHVV DQG / LV WKH FKDQQHO OHQJWK ,Q (TQf WKH VXUIDFH UHFRPELQDWLRQ UDWH 5VV FDQ EH LQGHSHQGHQWO\ FRQWUROOHG E\ 9*% DQG 9%( 9*% FRQWUROV WKH VXUIDFH EDQGEHQGLQJ FRQGLWLRQ DQG FDQ EH UHODWHG WR VXUIDFH EDQGEHQGLQJ SRWHQWLDO ?_V XVLQJ (TQf 9%( GHWHUPLQHV WKH VHSDUDWLRQ EHWZHHQ WKH TXDVL)HUPL SRWHQWLDOV IRU HOHFWURQ )1 DQG KROH )S DQG LV UHODWHG E\ 9%( )Q )SfT $V VKRZQ LQ (TQVf DQG f WKH VXUIDFH EDQGEHQGLQJ DQG TXDVL )HUPL OHYHO VSOLWWLQJ FRQWURO WKH VXUIDFH HOHFWURQ DQG KROH FRQFHQWUDWLRQV ([SUHVVLQJ 5VV 1,7\f}5 VHII ZKHUH 1,7 LV WKH DUHDO GHQVLW\ RI WKH LQWHUIDFH WUDSV DQG 5VHII LV WKH HIIHFWLYH UHFRPELQDWLRQ UDWH ,%B& FDQ EH IXUWKHU H[SUHVVHG DV U/f§\V ,%LF T Z 1,7\f 5VBHII G\ f \' ,W LV UHDVRQDEOH WR DVVXPH WKDW WKH VXUIDFH EDQG EHQGLQJ ?_Vf DQG WKH TXDVL)HUPL SRWHQWLDOV )1 DQG )Sf DUH SRVLWLRQLQGHSHQGHQW IRU VPDOO FKDQQHO FXUUHQW DQG ORQJ PLQRULW\ FDUULHU GLIIXVLRQ OHQJWK /'LII /f 7KHQ 5VHII FDQ EH WUHDWHG DV D FRQVWDQW DQG LQGHSHQGHQW RI SRVLWLRQ DQG (TQf EHFRPHV %& T : 5 6HII /\V 1LW \f G\ \' f

PAGE 72

&ROOHFWRU (PLWWHU )LJXUH Df &URVVVHFWLRQDO YLHZ RI WKH Q0267 WUDQVLVWRU XVHG LQ WKH 6('&,9 SURILOLQJ FRQILJXUDWLRQ Ef %DVH UHFRPELQDWLRQ FXUUHQW YHUVXV GUDLQ ELDV ,[ n 9'; RU ,E f 9&% IRU ILYH LQWHUIDFH WUDS GLVWULEXWLRQV &XUYH'UDLQ &KDQQHO 6RXUFHf DUH &XUYH f &XUYH QRQXQLIRUP f &XUYH FRQVWDQW f &XUYH QRQXQLIRUP f DQG &XUYH QRQXQLIRUP QRQXQLIRUP f

PAGE 73

)URP (TQf WKH GHFUHDVH LQ ,%& IRU DQ LQFUHPHQWDO LQFUHDVH LQ \' FDQ EH H[SUHVVHG DV $AELF
PAGE 74

RI WKH VSDWLDO GLVWULEXWLRQ RI WKH LQWHUIDFH WUDS LQ WKH UHJLRQ QHDU WKH GUDLQfV PHWDOOXUJLFDO SQ WUDQVLWLRQ ERXQGDU\ 7R REWDLQ D TXDOLWDWLYH YDOXH IRU 1,7 D PRGHO IRU 5VV QHHGV WR EH DVVXPHG $QDO\WLFDO PRGHOV IRU 5VV FDQ EH D VLQJOH HQHUJ\OHYHO 4,7 > @ RU D GLVWULEXWHG ',7 WKDW LV XQLIRUPO\ GLVWULEXWHG LQ HQHUJ\ QHDU WKH PLGGOH RI WKH VLOLFRQ HQHUJ\ JDS > @ 7KH FRQYHUVLRQ RI 9'; WR \SRVLWLRQ WR JLYH WKH \GHSHQGHQFH RI 17 FDQ EH REWDLQHG XVLQJ D 0267 GHYLFH VLPXODWRU ZLWK NQRZQ GRSLQJ SURILOH DQG GHYLFHfV VWUXFWXUH +RZHYHU DQ DQDO\WLFDO DEUXSW QS MXQFWLRQ DSSUR[LPDWLRQ >@ JLYHQ EHORZ FDQ EH XVHG DV D ILUVW RUGHU HVWLPDWLRQ WR UHODWH WKH GUDLQ VXUIDFH VSDFHFKDUJH OD\HU WKLFNQHVV WR WKH GUDLQ YROWDJH W f 9';BEL9'6f§OUVf @ f ZKHUH H6L LV WKH VLOLFRQ GLHOHFWULF FRQVWDQW 9';BEL N%7TfORJH1''1$$Qcf LV WKH EXLOWLQ SRWHQWLDO RI WKH GUDLQ QS MXQFWLRQ DW WKH LQWHUIDFH 7KH VLOLFRQ VXUIDFH HQHUJ\ EDQG EHQGLQJ LQ WKH [GLUHFWLRQ DW WKH GUDLQ ?`V GHSHQGV RQ WKH DSSOLHG JDWH YROWDJH DQG FDQ EH GHWHUPLQHG IURP WKH RQHGLPHQVLRQDO HOHFWURVWDWLF IRUPXOD RI D 026 FDSDFLWRU 1GG LV WKH GRQRU FRQFHQWUDWLRQ LQ WKH KHDYLO\GRSHG GUDLQ DQG 1$$ LV WKH DFFHSWRU FRQFHQWUDWLRQ LQ WKH SEDVH 1RWH WKDW LQ WKH FDVH RI 0267 ZLWK D OLJKWO\GRSHGGUDLQ VWUXFWXUH WKH \RA'[f HVWLPDWHG XVLQJ (TQf ZLOO EH ODUJHU WKDQ WKH DFWXDO YDOXH DV WKH YROWDJH GURS DFURVV WKH ORZO\GRSHG GUDLQ UHJLRQ LV QRW FRQVLGHUHG E\ WKLV VLPSOH DSSUR[LPDWLRQ %DVHG RQ (TQf TXDOLWDWLYH LQWHUSUHWDWLRQV RQ WKH VSDWLDO GLVWULEXWLRQ RI WKH LQWHUIDFH WUDSV FDQ EH GHGXFHG )RU H[DPSOH LI WKHUH LV D QRQXQLIRUP GLVWULEXWLRQ RI

PAGE 75

LQWHUIDFH WUDSV DORQJ WKH FKDQQHO DQG QHDU WR WKH GUDLQ UHJLRQ RI D Q0267 WKH PDJQLWXGH RI ,%& GHFUHDVHV ZKHQ WKH HOHFWURQLF ERXQGDU\ RI WKH VXUIDFH GUDLQERG\ VSDFHFKDUJHOD\HU H[WHQGV WRZDUGV WKH FHQWHU RI WKH FKDQQHO E\ D SRVLWLYH 9'; 7KH GHFUHDVH LQ ,%B & LV FDXVHG E\ WKH GHFUHDVHG QXPEHU RI LQWHUIDFH WUDSV DYDLODEOH IRU UHFRPELQDWLRQ LQ WKH SRUWLRQ RI FKDQQHO PRGXODWHG E\ 9'; )LJXUH Ef VKRZV WKH ILYH YDULDWLRQV RI WKH ,% FXUUHQW ZLWK 9'; IRU ILYH SRVVLEOH 1,7 VSDWLDO GLVWULEXWLRQV &XUYH f LV WKH ,% FKDUDFWHULVWLF IRU D IUHVK XQVWUHVVHG Q0267 )RU ORZ DQG XQLIRUP 1LW GLVWULEXWLRQ ,% GHFUHDVHV ZLWK D VPDOO FRQVWDQW VORSH ZKHQ 9'; LQFUHDVHV &XUYH f VKRZV D QRQXQLIRUP 1,7 GLVWULEXWLRQ ZLWKLQ WKH GUDLQERG\ VXUIDFH VSDFHFKDUJH OD\HU LH \ \' DQG XQLIRUP 1,7 GLVWULEXWLRQ DORQJ WKH FKDQQHO &XUYH f H[KLELW D FRQVWDQW $,%J9'; VORSH DQG KHQFH LQGLFDWHV D XQLIRUP 17 GLVWULEXWLRQ DORQJ WKH FKDQQHO )RU 9'; WKH FRQVWDQW ,% DOVR VXJJHVWV XQLIRUP 1,7 GLVWULEXWLRQ LQ \ \' &XUYH f VKRZV WKDW ,% GHFUHDVHV ZLWK D YDU\LQJ VORSH IRU 9'; 9 DQG WKXV LQGLFDWHV QRQXQLIRUP 1,7 GLVWULEXWLRQ $ XQLIRUP 1,7 GLVWULEXWLRQ LV LPSOLHG LQ \ \' E\ XVLQJ WKH VDPH LQWHUSUHWDWLRQV JLYHQ DERYH &XUYH f LOOXVWUDWHV WKH FDVH IRU D QRQXQLIRUP 1,7 GLVWULEXWLRQ DORQJ WKH FKDQQHO DQG DOVR ZLWKLQ WKH GUDLQ VXUIDFH VSDFH FKDUJHOD\HU ([SHULPHQW DQG 5HVXOWV ,QGXVWULDO QFKDQQHO 0267 LV XVHG LQ WKH GHPRQVWUDWLRQ RI WKLV '&,9 3URILOLQJ WHFKQLTXH 7KH Q0267 LV IDEULFDWHG LQ D SZHOO ZLWK QHSLOD\HUSVXEVWUDWH DQG KDV GUDZQ FKDQQHO OHQJWK / _LP GUDZQ FKDQQHO ZLGWK : ILP JDWH R[LGH WKLFNQHVV [2[ $ DQG SZHOO GRSLQJ FRQFHQWUDWLRQ 1$$ [OFPa 7KH 6(

PAGE 76

'&,9 FRQILJXUDWLRQ LV XVHG WR LQMHFW PLQRULW\ FDUULHUV HOHFWURQVf LQWR WKH SERG\ IURP WKH IRUZDUGELDVHG VRXUFH MXQFWLRQ ,Q WKLV LOOXVWUDWLRQ ZH JHQHUDWH QRQXQLIRUP 1,7 E\ VWUHVVLQJ WKH Q0267 XQGHU &+( VWUHVV FRQGLWLRQ 9'6 9 9*6 9 DQG 9V[ 9f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f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

PAGE 77

9*%9f )LJXUH (IIHFW RI FKDQQHO KRW HOHFWURQ VWUHVV RQ WKH 'UDLQ(PLWWHU DQG 6RXUFH (PLWWHU '&,9 FKDUDFWHULVWLFV $OO PHDVXUHG DW 9%( 9 9&% 9 DQG &+( VWUHVV FRQGLWLRQV DUH 9'6 9 DQG 9*6 9 IRU PLQXWHV 7KH VROLG OLQH UHIHU WR WKH SRVWVWUHVVHG PHDVXUHPHQWV ZKLOH WKH GRWWHG OLQH LV IRU WKH XQVWUHVVHG PHDVXUHPHQW 7KH XQVWUHVVHG ,%9*% PHDVXUHPHQW LV VLPLODU IRU ERWK GUDLQ DQG VRXUFH MXQFWLRQV DQG KHQFH RQO\ RQH FXUYH LV VKRZQ

PAGE 78

)LJXUHV Df DQG Ef VKRZV WKH $,%9%( FXUYHV DW WKUHH GLIIHUHQW JDWH ELDVHV 9JE 9 9 DQG 9f IRU 6( DQG '(,E9EH PHDVXUHPHQW FRQILJXUDWLRQ UHVSHFWLYHO\ ,Q )LJXUH Df WKH LQFUHDVH LQ $,% IRU 9*% 9 LV SULPDULO\ FRQWULEXWHG E\ WKH VXUIDFH UHFRPELQDWLRQ DORQJ WKH FKDQQHO LH ,% & FRPSRQHQW 7KH VXUIDFH EDQGEHQGLQJ SRWHQWLDO \V DW 9*% 9 LV FRPSXWHG WR EH HTXDO WR 9 ZKLFK LV FORVH WR WKH IODWEDQG FRQGLWLRQ 9*% 9f REWDLQHG E\ D KLJKIUHTXHQF\ &J9* PHDVXUHPHQW RQ D _LP[ILP FDSDFLWRU $W ORZ 9%( ELDV LW LV REVHUYHG WKH LGHDOLW\ IDFWRU Q LV DERXW 7KLV VXJJHVWV WKDW WKH KROH VXUIDFH FRQFHQWUDWLRQ 3V LV JUHDWHU WKDQ WKH VXUIDFH HOHFWURQ FRQFHQWUDWLRQ 1V ZKLFK UHVXOWV LQ WKH VXUIDFH UHFRPELQDWLRQ UDWH WR EH GHSHQGHQW RQ 9%( 7KXV $,% rF ,% & RF H[ST9%(N%7f $W KLJK 9%( ELDV LW LV REVHUYHG WKDW WKHUH LV D FKDQJH RI WKH LGHDOLW\ IDFWRU IURP Q O WR Q 7KLV LV FRQVLVWHQW ZLWK 5VV H[ST9%(N%7f IRU 3V 1V DV GLVFXVVHG LQ &KDSWHU 7KH WUDQVLWLRQ IURP Q O WR Q RFFXUV DW DERXW 9 DQG DJUHHV UHDVRQDEO\ ZHOO ZLWK WKH WKHRUHWLFDO YDOXH SUHGLFWHG XVLQJ D VLPSOH VLQJOH HQHUJ\OHYHO LQWHUIDFH WUDS PRGHO 9%( >9) f§ \UV 9*% f§9f @ f > N%7 Tf ORJH 1Af 9@ >9 9@ 9 ,Q )LJXUH Ef WKH LQFUHDVH LQ $,% IRU WKUHH JDWH ELDV YDOXHV LV DWWULEXWHG WR WKH VXUIDFH UHFRPELQDWLRQ ZLWKLQ WKH GUDLQERG\ VXUIDFH VSDFHFKDUJHOD\HU 7KLV LV VXSSRUWHG E\

PAGE 79

)LJXUH (IIHFW RI 9*% RQ WKH $,%9%( FKDUDFWHULVWLF IRU Df 6RXUFH(PLWWHU %-7 DQG Ef IRU 'UDLQ(PLWWHU %-7 $OO PHDVXUHG DW 9&% 9 DQG .

PAGE 80

WKH H[SHULPHQWDOO\ REVHUYHG Q LGHDOLW\ IDFWRU REWDLQHG E\ ILWWLQJ $,%B'(9%( WR D VLQJOH H[SRQHQWLDO IXQFWLRQ >$,% $,%2H[ST9%(QN%7f@ )LJXUHV Df DQG Ef VKRZ WKH SRVW DQG SUHVWUHVVHG 6RXUFH(PLWWHU ,%9&% H[SHULPHQWDO UHVXOWV $,%9&% DQG >$,%fG9&%@ UHVSHFWLYHO\ ,Q )LJXUH Df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f ZKHQ 9&% 9 LQGLFDWHV KROH FXUUHQW IURP LQWHUEDQG LPSDFW JHQHUDWLRQ E\ WKH LQMHFWHG HOHFWURQV ZKLFK DUH DFFHOHUDWHG E\ WKH KLJK HOHFWULF ILHOG LQ WKH UHYHUHELDVHG a9f GUDLQ MXQFWLRQ VSDFH FKDUJHOD\HU 7KXV WKLV OLPLWV WKH UDQJH RI 9&% WKDW FDQ EH DSSOLHG WR H[WHQG WKH GUDLQERG\ VSDFHFKDUJH OD\HU RXW LQWR WKH FKDQQHO ZLWKRXW FDXVLQJ DQ\ DGGLWLRQDO VWUHVV RQ WKH GHYLFH GXULQJ SURILOLQJ PHDVXUHPHQW 7R H[WUDFW WKH VWUHVVJHQHUDWHG 17 IURP WKH PHDVXUHG SRVW DQG SUHVWUHVVHG ,%9&% ZH SHUIRUP WKH IROORZLQJ DQDO\VLV RQ WKH GDWD 7KH GLIIHUHQFH UHVXOW LV VKRZQ LQ )LJXUH Ef $,E V > ,% 9%( 9 WVWU 2PLQVf f§,% 9%( 9 WVWU f @ >,% 9%( 9 WVWU PLQVf ,% 9%( 9 WVWU f @ f

PAGE 81

9&%9f &+$11(/ \ XUQf O , &f )LJXUH 6WUHVVDQGPHDVXUH 6$0f GDWD RI EDVH UHFRPELQDWLRQ FXUUHQW YDULDWLRQ ZLWK GUDLQ ELDV XVLQJ IRUZDUGELDVHG Q VRXUFHSERG\ MXQFWLRQ DV WKH PLQRULW\ FDUULHU HOHFWURQf HPLWWHU &+( VWUHVV DW 9'6 9 9*6 9 DQG 9V[ 9 IRU PLQXWHV 0HDVXUHG DW 9%( 9 9*% 9 DQG 9&% 9 Df 5DZ GDWD VKRZLQJ GHFUHDVH DW 9'; !a 9 GXH WR KROH JHQHUDWHG E\ HOHFWURQ LPSDFW Ef &+( JHQHUDWHG H[FHVV EDVH FXUUHQW $,% DQG $,Ef9FE YHUVXV 9&% Ff (VWLPDWHG LQWHUIDFH WUDS SRVLWLRQ IURP WKH Q GUDLQ MXQFWLRQ ERXQGDU\ ,S$9 FRUUHVSRQGV WR 17 WUDSVFPf

PAGE 82

,Q (TQf WKH ILUVW VXEWUDFWLRQ XVHG LV WR HOLPLQDWH WKH FRQWULEXWLRQ IURP WKH GUDLQ VXUIDFH UHFRPELQDWLRQ ,%6 FRPSRQHQW 7KH VHFRQG VXEWUDFWLRQ XVHG LV WR FDQFHO RXW WKH LPSDFWJHQHUDWHG KROH FXUUHQW FRPSRQHQW 7KH ILQDO VXEWUDFWLRQ JLYHV WKH VWUHVVn JHQHUDWHG EDVH FXUUHQW $,% DV VKRZQ LQ )LJXUH Ef ,W LV REVHUYHG WKDW WKHUH LV D FRQWLQXRXV ULVH RI WKH ,% aS$f IURP WKH VRXUFH WRZDUGV WKH GUDLQ aS$f ZKLFK LV H[SHFWHG IURP WKH FKDQQHO KRW HOHFWURQ VWUHVV 7KLV $,%9&% GDWD LV WKHQ OHDVWVTXDUH ILWWHG WR REWDLQ WKH VORSH RI $,%f9&% ZKLFK LV VKRZQ LQ )LJXUH Ef (YLGHQWO\ WKHUH LV D FKDQJH LQ VORSH LQ ,% DV 9&% YDULHV ZKLFK LQGLFDWHV D QRQXQLIRUP GLVWULEXWLRQ RI VWUHVVJHQHUDWHG LQWHUIDFH WUDSV 7KH UHVXOWV TXDOLWDWLYHO\ VKRZV WKDW WKH VWUHVVn JHQHUDWHG LQWHUIDFH WUDS VWDUWV WR LQFUHDVH DW DERXW _LP IURP WKH Q GUDLQ MXQFWLRQ ERXQGDU\ $,%f9&% a S$9f UHDFKLQJ DERXW IRXU WLPHV aS$9f DW SP IURP WKH Q GUDLQ MXQFWLRQ ERXQGDU\ 7KLV LV FRQVLVWHQW ZLWK WKH VSDWLDO YDULDWLRQ RI WKH HOHFWULF ILHOG ZKHUH HOHFWULF ILHOG VWUHQJWK LQFUHDVHV DIWHU WKH GHSOHWLRQ SRLQW \ \9'6DWf! DQG UHDFKHV D PD[LPXP DW WKH GUDLQ PHWDOOXUJLFDO MXQFWLRQ ,Q VXPPDU\ D VLPSOH DQG UHOLDEOH H[WHQVLRQ RI WKH '&,9 WHFKQLTXH LV GHPRQVWUDWHG WR SURILOH WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDS GHQVLWLHV DORQJ WKH FKDQQHO RI 026 WUDQVLVWRUV

PAGE 83

&+$37(5 /,1($5 5('8&7,21 2) '5$,1 &855(17 :,7+ ,1&5($6,1* ,17(5)$&( 5(&20%,1$7,21 ,1 1026 75$16,67256 675(66(' %< &+$11(/ +27 (/(&7521 ,QWURGXFWLRQ 5HFHQWO\ 1HXJURVFKHO DQG 6DK SURSRVHG D VLPSOH DQG KLJKO\ VHQVLWLYH PHWKRG WKH 'LUHFW&XUUHQW &XUUHQW9ROWDJH '&,9f PHWKRG WR PRQLWRU WKH GHJUDGDWLRQ RI VLOLFRQ 026 WUDQVLVWRUV 0267Vf XQGHU FKDQQHO DQG VXEVWUDWH KRW FDUULHU VWUHVVHV >@ 7KH '&,9 PHWKRG XVHV WKH JDWHYROWDJHFRQWUROOHG LQWHUIDFLDO UHFRPELQDWLRQ FXUUHQW ,% EDVH RU ERG\ FXUUHQWf WR PRQLWRU WKH JHQHUDWLRQ UDWH RI LQWHUIDFH WUDSV DQG WKH FKDUJLQJ UDWH RI R[LGH WUDSV LQ 0267V +RZHYHU WKH WUDGLWLRQDO WUDQVLVWRU UHOLDELOLW\ PHDVXUHPHQWV PRQLWRU WKH FKDQJHV RI WKH VDWXUDWLRQ GUDLQ FXUUHQW $,'BVDWf WKUHVKROG JDWH YROWDJH $9*7f DQG VXEWKUHVKROG YROWDJH VZLQJ $6f > @ ,Q WKLV FKDSWHU ZH GHPRQVWUDWH WKH DQWLFLSDWHG OLQHDU GHSHQGHQFLHV RI WKHVH WUDGLWLRQDO FKDUDFWHULVWLFV RQ WKH '&,9 EDVHERG\ UHFRPELQDWLRQ FXUUHQW LH $,'VDW $9*7 $6 $,E ([SHULPHQWV 7KH QFKDQQHO 0267V Q0267Vf XVHG KDYH D GUDZQ FKDQQHO ZLGWKOHQJWK DVSHFW UDWLR RI /$9 SP_OP D $ JDWH R[LGH DQG VHSDUDWH VRXUFH GUDLQ JDWH ERG\ FRQWDFW SDGV 'XULQJ WKH 6WUHVVDQG0HDVXUH 6$0f H[SHULPHQW UXQV WKH Q0267V ZHUH VWUHVVHG E\ FKDQQHO KRW HOHFWURQV &+(Vf LQ WKH IRUZDUG FRQILJXUDWLRQ LH

PAGE 84

QGUDLQSERG\ MXQFWLRQ LV UHYHUVHELDVHGf ZKLOH WKHLU HOHFWULFDO FKDUDFWHULVWLFV ZHUH PHDVXUHG DIWHU HDFK VWUHVV GXUDWLRQ LQ WKH UHYHUVH FRQILJXUDWLRQ LH QVRXUFHSERG\ MXQFWLRQ LV UHYHUVHELDVHG WR PHDVXUH $,VVDW $9*7 DQG $6 DQG WKH QGUDLQSERG\ LV IRUZDUGELDVHG WR PHDVXUH $,% LQ WKH '&,9 PRGH ZKLFK LV NQRZQ DV WKH 'UDLQ(PLWWHU '&,9 RU '('&,9 f 7KLV LQWHUFKDQJH RI VRXUFH DQG GUDLQ GXULQJ PHDVXUHPHQWV LV WUDGLWLRQDOO\ HPSOR\HG EHFDXVH WKH JHQHUDWHG LQWHUIDFH WUDSV DQG FKDUJHG R[LGH WUDSV DUH ORFDWHG RYHU WKH GUDLQMXQFWLRQ VSDFHFKDUJH UHJLRQ 7ZR ELDV YROWDJH FRQGLWLRQV DUH XVHG GXULQJ WKH &+( VWUHVV GXUDWLRQ Lf WKH WUDGLWLRQDO 9'6 9*6 ZKLFK JLYHV WKH PD[LPXP ERG\ FXUUHQW ,[6WUHVV! DQFr LLf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rV GXH WR WKH VWUHVVJHQHUDWHG 6L6L LQWHUIDFH WUDS RYHU WKH GUDLQ MXQFWLRQ VSDFHFKDUJH UHJLRQ > @ 7KH QHJOLJLEOH 9* VKLIW RI WKH ,% LQGLFDWHV

PAGE 85

)LJXUH 'UDLQ(PLWWHU '('&,9 FKDUDFWHULVWLFV RI D Q0267 VWUHVVHG E\ FKDQQHO KRW HOHFWURQV

PAGE 86

QHJOLJLEOH FKDUJLQJ RI WKH R[LGH WUDSV RYHU WKH GUDLQMXQFWLRQ VSDFHFKDUJH UHJLRQ EHFDXVH WKH &+(V GR QRW KDYH VXIILFLHQW NLQHWLF HQHUJ\ .( T<'69'6BVDWf T><'V9*9*7f@ H9 WR VXUPRXQW WKH H9 6L6L EDUULHU 7KH &+(V DOVR GR QRW KDYH VXIILFLHQW NLQHWLF WR FUHDWH VHFRQGDU\ KRW KROHV YLD LQWHUEDQG LPSDFW HOHFWURQKROH SDLU JHQHUDWLRQ WR VXUPRXQW WKH H9 6L6L KROH EDUULHU ZKLFK ZRXOG UHTXLUH D PLQLPXP HQHUJ\ RI T9'69'6VDWf _!%BKROH (JDS6M f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f&+Q9*6 9*7f f ZH JHW AfA'VDWA A'VDW2 f§ $9JWf 9JVf§9JWf f ZKHUH WKH VXEVFULSW GHQRWHV WKH SUHVWUHVV YDOXH (TQf DVVXPHV WKDW 9*9*7f } $9JW 6LQFH $9*7 T$127 $1,7f&; DQG ,% 1,7 > @ ZH FDQ LPPHGLDWHO\ REWDLQ $,'VDW m $9*7 r $,%

PAGE 87

9G 9J 9f )LJXUH )RUZDUG DQG UHYHUVH ,'VDW FKDUDFWHULVWLF RI D Q0267 VWUHVVHG E\ FKDQQHO KRW HOHFWURQV 'LIIHUHQFH EHWZHHQ WKH SRVWVWUHVV IRUZDUG )f DQG UHYHUVH 5f FKDUDFWHULVWLFV UHYHDO WKH DV\PPHWULFDO QDWXUH RI FKDQQHO KRW HOHFWURQ VWUHVV WVWUHVV NVHF 3UHVWUHVV IRUZDUG DQG UHYHUVH ,'VDW FKDUDFWHULVWLFV DUH LGHQWLFDO

PAGE 88

,' $f 9*6 9f )LJXUH )RUZDUG DQG UHYHUVH ,'9* FKDUDFWHULVWLF RI D Q0267 VWUHVVHG E\ FKDQQHO KRW HOHFWURQV 'LIIHUHQFH EHWZHHQ WKH SRVWVWUHVV IRUZDUG )f DQG UHYHUVH 5f FKDUDFWHULVWLFV UHYHDO WKH DV\PPHWULFDO QDWXUH RI FKDQQHO KRW HOHFWURQ VWUHVV WVWUHVV NVHF )RU FODULW\ RQO\ WKH XQVWUHVV IRUZDUG ,'9* FKDUDFWHULVWLF LV VKRZQ $OO PHDVXUHG DW 9'6 9

PAGE 89

'VDWn 'VDW2 r &7 LFU a b$, %'(SHDN )LJXUH 'HSHQGHQFH RI $,'BVDW,'VDW $9*7 DQG $6 RQ $,%SHDN,%BSHDN LQ ALP VLOLFRQ Q0267V VWUHVVHG E\ FKDQQHO KRW HOHFWURQV $ 9GV9JV $9GV 9 9JV 9 $œ669 fs$9

PAGE 90

7KH VWUHVVLQGXFHG ZLGHQLQJ RI WKH VXEWKUHVKROG VZLQJ $6 LV DOVR VKRZQ LQ )LJXUH ,W LQFUHDVHV LQLWLDOO\ DV $6 $,% WKHQ ULVHV IDVWHU DQG ILQDOO\ DSSURDFKHV DJDLQ $6 RF $,% 7KLV LV FRQVLVWHQW ZLWK WKH VLPSOH VXEWKUHVKROG VORSH H[SUHVVLRQ 6 N7Tf >O&LW&Vf&R[@ f ZKHUH &LW T'>7 DQG 1,7 M ,7(7fI7(7fG(7 >@ ZKLFK LQGLFDWHV DQ LQLWLDO DQG ILQDO OLQHDU GHSHQGHQFH $6 $&LW $1,7 m $,% ZKLOH WKH VXSHUOLQHDU ULVH RI $6 LQ WKH PLGUDQJH RI VWUHVV LV FDXVHG E\ WKH IDVWHU FKDQJH RI WKH VXUIDFH SRWHQWLDO ZLWK $9*7 UHODWLYH WR WKH HQHUJ\ GLVWULEXWLRQ RI 'MMI(Mf ,Q VXPPDU\ DQ H[SHULPHQWDO SURRI LV JLYHQ RI WKH WKHRUHWLFDOO\ DQWLFLSDWHG OLQHDU GHSHQGHQFH RI $,%BSHDN RQ WKH WUDGLWLRQDOO\ PRQLWRUHG WUDQVLVWRU UHOLDELOLW\ FKDUDFWHULVWLFV $,'BVDW DQG $9*7 7KLV FRUUHODWLRQ SURYLGHV WKH EDVLV IRU XVLQJ WKH '&,9 PHWKRG WR PRQLWRU 026 WUDQVLVWRU GHJUDGDWLRQ

PAGE 91

&+$37(5 ,17(5)$&( 75$36 *(1(5$7,21 02'(/ ,QWURGXFWLRQ $V VKRZQ LQ &KDSWHUV DQG LQWHUIDFH WUDSV DUH JHQHUDWHG ZKHQ 026 WUDQVLVWRUV DUH VWUHVVHG XQGHU FKDQQHO KRW FDUULHU VWUHVV FRQGLWLRQV 7KHVH LQWHUIDFH WUDSV DUH HOHFWURQKROH UHFRPELQDWLRQJHQHUDWLRQ VLWHV DQG WKRXJKW WR EH GDQJOLQJ VLOLFRQ DQG R[\JHQ ERQGV 6L 6L2f >@ FUHDWHG E\ HQHUJHWLF HOHFWURQV DQG KROHV EUHDNLQJ WKH ZHDN LQWULQVLF 6L6L 6L26Lf RU LPSXULW\ 6L+ 6L2+f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

PAGE 92

EDVHG RQ D SK\VLFDO ERQGEUHDNLQJ PHFKDQLVP DQG FRQVLGHUV WKH ERQGEUHDNLQJ HIILFLHQF\ RI WKH KRW HOHFWURQV DQG KROHV DORQJ WKH VXUIDFH FKDQQHO RQ R[LGL]HG VLOLFRQ ,W LGHQWLILHV WKH SULPDU\ DQG GRPLQDQW KRW FDUULHUV HOHFWURQV RU KROHVf WKDW DUH LQYROYHG LQ WKH LQWHUIDFH WUDS JHQHUDWLRQ SURFHVV ,Q WKLV FKDSWHU WKH DQDO\WLFDO LQWHUIDFH WUDS JHQHUDWLRQ WKHRU\ LV SUHVHQWHG DQG WKH YDOLGLW\ RI WKH PRGHO LV VXSSOHPHQWHG E\ H[SHULPHQWDO YHULILFDWLRQ 7KHRU\ &RQVLGHU D QFKDQQHO 026 WUDQVLVWRU ELDVHG LQ WKH IODWEDQG FRQGLWLRQ VKRZQ LQ )LJXUH Df WKH GUDLQ QS MXQFWLRQ LV UHYHUVH ELDVHG E\ 95 ZKLFK UHVXOWV LQ D ZLGHU VSDFHFKDUJHOD\HU 6&/f 7KH FRUUHVSRQGLQJ HQHUJ\ EDQG GLDJUDP LV GHSLFWHG LQ )LJXUH Ef 7KH GUDLQ MXQFWLRQ VSDFHFKDUJHOD\HU FDQ DFFHOHUDWH HOHFWURQV LQMHFWHG DW WKH ORZILHOG HGJH /)(f RI WKH GUDLQ 6&/ DQG KROHV LQMHFWHG DW WKH KLJKILHOG HGJH +)(f RI WKH GUDLQ 6&/ WR KLJKHU NLQHWLF HQHUJ\ ZKHQ WKH HOHFWURQV RU KROHV WUDQVLW DFURVV WKH 6&/ :KHQ WKH DFFHOHUDWLQJ HOHFWURQV DQG KROHV DFTXLUH VXIILFLHQW HQHUJ\ JUHDWHU WKDQ WKH ERQGEUHDNLQJ HQHUJ\ T9%%f WKH\ FDQ EUHDN WKH K\GURJHQDWHG LQWHUIDFLDO ERQGV 6L+ DQG 6L2+f WR UHOHDVH WKH K\GURJHQ DQG FUHDWH QHZ LQWHUIDFH WUDSV %DVHG RQ 6DKfV LQWHUIDFH WUDS JHQHUDWLRQ PRGHO WKH UDWH RI LQWHUIDFH WUDS JHQHUDWLRQ G17GWf LV SURSRUWLRQDO WR WKH LQMHFWHG FDUULHU IOX[ GHQVLW\ -MTf WKH SUREDELOLW\ RI VXUYLYLQJ D VFDWWHULQJ DQG WKH SUREDELOLW\ RI ERQGEUHDNLQJ 7KXV WKH LQWHUIDFH WUDS JHQHUDWLRQ UDWH LV JLYHQ E\ WKH SURGXFW RI WKHVH WKUHH WHUPV

PAGE 93

)LJXUH &URVVVHFWLRQDO YLHZ RI D UHYHUVH ELDVHG SQ MXQFWLRQ FRYHUHG E\ 6L VKRZLQJ WKUHH GRPLQDQW ERQGEUHDNLQJ SDWKZD\V WR JHQHUDWH LQWHUIDFH WUDSV Df &URVVVHFWLRQDO YLHZ Ef 7UDQVLWLRQ HQHUJ\ EDQG GLDJUDP DORQJ WKH R[LGHVLOLFRQ LQWHUIDFH %RQG EUHDNLQJ LV PDUNHG VWDU r DQG YHUWLFDO GDVK OLQH 2SHQ WULDQJOHV DUH LQWHUIDFH WUDSV )LOOHG WULDQJOHV DUH 6L+ DQG 6L2+ 'RWV DUH HOHFWURQV DQG FLUFOHV DUH KROHV (OHFWURQ DQG KROH PRWLRQV DUH PDUNHG E\ DUURZV ZLWK VROLG DQG RSHQ WLSV UHVSHFWLYHO\ $GDSWHG IURP 6DK > S @

PAGE 94

G1,7 -M A%% \G%%f$@ LV WKH IUDFWLRQ RI WKH LQMHFWHG FDUULHU ZKLFK KDV HVFDSHG VFDWWHULQJ DQG QRW EURNHQ ERQG ZKHQ WKH FDUULHU WUDQVLWV IURP \ \G%%f WR \ \ G\;%%f LV WKH IUDFWLRQ RI ERQGEUHDNLQJ HYHQWV ZLWKLQ WKH G\ UHJLRQ 7KH FRPELQHG PHDQIUHHSDWK $ FRQWDLQV WKUHH FRPSRQHQWV WKH VXUIDFH VFDWWHULQJ $V ZKLFK LQFOXGHV WKH SKRQRQ LPSXULW\ HOHFWURQHOHFWURQ DQG HOHFWURQKROH VFDWWHULQJ HYHQWV WKH LQWHUEDQG LPSDFWJHQHUDWLRQ RI HOHFWURQKROH SDLU SQ DQG WKH ERQGEUHDNLQJ E\ WKH HQHUJHWLF FDUULHU "A%% $ $V 31 ;%% f % ƒ $EE f 6LQFH WKH ERQGEUHDNLQJ HYHQW RFFXUV LQIUHTXHQWO\ LW LV DQWLFLSDWHG WKDW $%% } $31 ;V A A 6ROYLQJ (TQf DQG XVLQJ WKH DSSUR[LPDWLRQ
PAGE 95

n,7 U G177 U-L L f§ / GW / T H[S A%% r ƒ A%% +LWR f f f A%% ZKHUH UL,7 H[SG%%;f LV WKH QRUPDOL]HG LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ $QDO\WLFDO H[SUHVVLRQV IRU YROWDJH GHSHQGHQFH RI WKH ERQGEUHDNLQJ GLVWDQFH G%% FDQ EH GHULYHG E\ XVLQJ WKH GHSOHWLRQ DSSUR[LPDWLRQ RI DQ DEUXSW QS MXQFWLRQ DQG E\ WUDFLQJ WKH WKUHH ERQGEUHDNLQJ SDWKZD\V LOOXVWUDWHG LQ )LJXUH RI WKH HQHUJHWLF HOHFWURQV DQG KROHV )LUVW OHW FRQVLGHU WKH HOHFWURQ LQMHFWLRQ IURP WKH ORZILHOG HGJH RI WKH GUDLQ VSDFH FKDUJH OD\HU WKH /)(L SDWKZD\ VKRZQ LQ )LJXUH f WKH DFFHOHUDWLQJ HOHFWURQ LQ WKH GUDLQ VSDFHFKDUJH OD\HU FDQ EUHDN DQ LQWHUIDFLDO ERQG ZKHQ LW KDV NLQHWLF HQHUJ\ JUHDWHU WKDQ WKH ERQGEUHDNLQJ HQHUJ\ T9%% 7KXV G%% FDQ EH H[SUHVVHG DV IROORZV nW%% T1Af 9%% f f ZKHUH LV WKH VLOLFRQ GLHOHFWULF FRQVWDQW T LV WKH HOHFWURQLF FKDUJH DQG 1$$ LV WKH DFFHSWRU GRSLQJ FRQFHQWUDWLRQ )RU WKLV /)(L LQWHUIDFH WUDS JHQHUDWLRQ SDWKZD\ WKH ERQGEUHDNLQJ GLVWDQFH LV LQGHSHQGHQW RI WKH DSSOLHG YROWDJH DQG RQO\ GHSHQGV RQ WKH DFFHSWRU GRSLQJ FRQFHQWUDWLRQ 1$$ DQG WKH WKUHVKROG ERQGEUHDNLQJ HQHUJ\ T9%% 7KLV LPSOLHV WKDW WKH HIILFLHQF\ RI LQWHUIDFH WUDS JHQHUDWLRQ E\ WKHVH HQHUJHWLF KRW HOHFWURQV LV D FRQVWDQW ZKHQ WKH DSSOLHG YROWDJH H[FHHGV WKH ERQGEUHDNLQJ HQHUJ\ DQG

PAGE 96

LV ]HUR LI WKH DSSOLHG YROWDJH LV EHORZ 7KH QRUPDOL]HG LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ IRU WKLV /)(L SDWKZD\ LV +LWR H[S >M eT1$$f ƒ?? 9%% f 1H[W FRQVLGHU WKH KROH LQMHFWLRQ SDWKZD\ IURP WKH KLJKILHOG GUDLQ HGJH ODEHOOHG +)(L LQ )LJXUH Df 7KH VRXUFHV RI WKHVH KROHV FDQ EH GXH WR WKHUPDO JHQHUDWLRQ RU IURP WKH LQWHUEDQGLPSDFW JHQHUDWLRQ SURFHVV LQLWLDWHG E\ WKH HQHUJHWLF HOHFWURQ ZKHQ WKH HOHFWURQ NLQHWLF HQHUJ\ .(HOHFWURQ LV JUHDWHU WKDQ WKH HOHFWURQKROH SDLU JHQHUDWLRQ WKUHVKROG (31 (*6L H9f >@ 7\SLFDOO\ WKH FRQFHQWUDWLRQ RI WKH (OHFWURQ ,PSDFW*HQHUDWHG (,+f KROHV LV KLJKHU WKDQ WKH WKHUPDOO\JHQHUDWHG KROH FRQFHQWUDWLRQ 7KHVH LPSDFWJHQHUDWHG KROHV FDQ EH EDFNLQMHFWHG LQWR WKH GUDLQ VSDFH FKDUJH OD\HU IURP WKH PD[LPXPILHOG HGJH DQG EUHDN WKH LQWHUIDFLDO ERQGV ZKHQ LWV HQHUJ\ H[FHHGV WKH ERQGEUHDNLQJ HQHUJ\ 7KH ERQGEUHDNLQJ GLVWDQFH IRU WKLV +)+L LQWHUIDFH WUDS JHQHUDWLRQ SDWKZD\ LV %%  eT1$$f >L 959%, L 9U9%L 9EE @ f 6LPLODUO\ WKH FRUUHVSRQGLQJ QRUPDOL]HG LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ LV JLYHQ E\ + LWR f§ H[S > W_ e f ]A@ & 959%L 959%, 9%% @ f ZKHUH 9%, LV WKH EXLOWLQ SRWHQWLDO RI WKH DEUXSW QS GUDLQ MXQFWLRQ &RPSDULQJ ZLWK (TQVf DQG f HTQVf DQG f UHYHDO WKDW WKLV +)+L LQWHUIDFH WUDS JHQHUDWLRQ SDWKZD\ KDV D YHU\ VWURQJ GHSHQGHQFH RQ WKH DSSOLHG VWUHVV YROWDJH 95 7KLV FKDUDFWHULVWLF VWURQJ YROWDJH GHSHQGHQFH RI G%% RU UM7 SURYLGHV D IHDWXUH IRU WKH WKHRU\ WR EH H[SHULPHQWDOO\ YHULILHG 7KLV ZLOO EH GHVFULEHG DQG HODERUDWH LQ WKH

PAGE 97

VXEVHTXHQW VHFWLRQ (TQVf DQG f DSSO\ RQO\ WR WKH GUDLQFXUUHQW UDQJH RI WKH 0267 ZKHUH 95 ,9' 9'VDWO ZKHUH 9'VDW LV WKH GUDLQ YROWDJH DW GUDLQ FXUUHQW VDWXUDWLRQ DQG PD\ DSSUR[LPDWHG E\ ,9* 9*7, 7KH 9%, PXVW DOVR EH PRGLILHG WR WDNH LQWR DFFRXQW RI WKH EXLOWLQ SRWHQWLDO EDUULHU KHLJKW GLIIHUHQFH EHWZHHQ WKH GUDLQERG\ DQG VRXUFHERG\ MXQFWLRQV )RU LGHQWLFDO GUDLQ DQG VRXUFH MXQFWLRQ 9%, ZKHQ WKH 0267fV FKDQQHO LV LQYHUWHG $QRWKHU SRVVLEOH SDWKZD\ IRU KROH WR JHQHUDWH LQWHUIDFH WUDSV LV YLD WKH ,QWHUEDQG 7XQQHOLQJ +ROH ,QMHFWLRQ ODEHOHG DV ,7+L LQ )LJXUH f 7KLV ,7+L SDWKZD\ ZLOO EHFRPH LPSRUWDQW LQ KHDYLO\ GRSHG QHPLWWHUSEDVH MXQFWLRQ RI DGYDQFHG VXEPLFURQ %-7V DQG IXWXUH JHQHUDWLRQ RI 0267V ZKHUH WKH ERG\ GRSLQJ FRQFHQWUDWLRQ LV DSSURDFKLQJ DQG H[FHHGLQJ FP $W WKLV KLJK FRQFHQWUDWLRQ YDOHQFH HOHFWURQV LQ WKH ERG\EDVHf UHJLRQ FDQ TXDQWXP PHFKDQLFDOO\ WXQQHO LQWR WKH HPSW\ VWDWHV RI WKH FRQGXFWLRQ EDQG LQ WKH TXDVLQHXWUDO GUDLQ HPLWWHUf UHJLRQ OHDYLQJ EHKLQG WKHUPDO KROHV 6LPLODUO\ WR WKH LQWHUEDQG LPSDFWJHQHUDWHG KROHV WKHVH LQWHUEDQG WXQQHOLQJ KROHV FDQ WKHQ DFFHOHUDWH LQ WKH KLJKILHOG 6&/ WR FUHDWH QHZ LQWHUIDFH WUDSV 7KH ERQGn EUHDNLQJ GLVWDQFH DVVRFLDWHG ZLWK WKLV ,7+L LQWHUIDFH WUDS JHQHUDWLRQ SDWKZD\ LV JLYHQ DV f%% ? eT1$$f >L 959%, 9VLB*DS 9)1 95 9%, 96L*DS Y )1 A)1 A%% f ,Q (TQf WKH YDOHQFHHOHFWURQ WXQQHOLQJ SURFHVV GHFUHDVHV WKH YROWDJH GHSHQGHQFH RI WKH ERQG EUHDNLQJ GLVWDQFH E\ 96L*DS 9)1f ZKHUH 96LB*DS LV WKH VLOLFRQ HQHUJ\

PAGE 98

JDS SRWHQWLDO DQG 9)1 LV WKH )HUPL OHYHO SRWHQWLDO DERYH WKH FRQGXFWLRQ EDQG HGJH RI WKH TXDVLQHXWUDO Q GUDLQHPLWWHU 7KH FRUUHVSRQGLQJ QRUPDOL]HG LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ LV WKHQ JLYHQ E\ :LWR H[3>O 6(T1Af ;@ >M959%,9VLB*DS9)1 959%L9VLB*DS9)1B 9%% f $V VKRZQ LQ (TQV f DQG f WKH LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ KDV D VWURQJ YROWDJH GHSHQGHQFH RQ WKH DSSOLHG VWUHVV YROWDJH ZKHQ WKH HOHFWURQ RU KROHf HQWHUV WKH VSDFHFKDUJHOD\HU IURP WKH KLJKILHOG HGJH 2Q WKH FRQWUDU\ LI WKH HOHFWURQ RU KROHf HQWHUV IURP WKH ORZILHOG HGJH RI WKH VSDFHFKDUJHOD\HU WKH LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ KDV QR YROWDJH GHSHQGHQFH DV VHHQ IURP (TQf +HQFH D YROWDJHGHSHQGHQF\ H[SHULPHQW FDQ EH GHVLJQHG WR FRUUHODWH WKH H[SHULPHQWDO UHVXOWV ZLWK WKH LQWHUIDFH WUDS JHQHUDWLRQ WKHRU\ SUHVHQWHG LQ WKLV VHFWLRQ ([SHULPHQW DQG 5HVXOWV )RXU VHWV RI 7LPHWR)DLOXUH 6$0 H[SHULPHQWV RQ VLOLFRQ Q0267 S0267 DQG QSQ%-7 ZHUH DQDO\]HG XVLQJ WKH QHZ LQWHUIDFH WUDS JHQHUDWLRQ WKHRU\ 7KH Q0267 GDWD DUH REWDLQHG IURP D FKDQQHOKRWHOHFWURQ &+(f VWUHVV H[SHULPHQW ZKHUH WKH Q0267 WHVW VWUXFWXUH ZDV IDEULFDWHG IURP D PLFURQ %L&026 WHFKQRORJ\ ZLWK _LP FKDQQHO OHQJWK SP FKDQQHO ZLGWK DQG $ WKLFN JDWH R[LGH 7KH S0267 GDWD DUH VLPLODUO\ REWDLQHG IURP D FKDQQHOKRWKROH &++f VWUHVV H[SHULPHQW ZKHUH WKH S0267 WHVW VWUXFWXUH ZDV IDEULFDWHG IURP D PLFURQ &026 WHFKQRORJ\ ZLWK _LP FKDQQHO OHQJWK _LP FKDQQHO ZLGWK DQG $ WKLFN JDWH R[LGH 7KH H[SHULPHQWDO

PAGE 99

GHWDLOV RI WKH Q0267 DQG S0267 6$0 GDWD ZLOO EH GHVFULEHG LQ &KDSWHU 7KH WZR VHWV RI QSQ%-7 GDWD DUH REWDLQHG IURP HPLWWHUEDVH UHYHUVHELDV VWUHVV H[SHULPHQW 7KH QSQ%-7V ZHUH IDEULFDWHG IURP D VXEPLFURQ %L&026 WHFKQRORJ\ ZLWK [O FP EDVH GRSLQJ FRQFHQWUDWLRQ DQG _LP[_LP HPLWWHU FRQWDFW DUHD 'XULQJ WKH HPLWWHUEDVH VWUHVV H[SHULPHQW WKH HPLWWHUEDVH MXQFWLRQ RI WKH QSQ%-7V ZDV UHYHUVHELDVHG DW 95 ZKLOH WKH FROOHFWRUEDVH MXQFWLRQ ZDV HLWKHU RSHQFLUFXLWHG 2&f RU IRUZDUGELDVHG )&f $ b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f ZKHUH 9% DQG 95 9'9'VDW 7KH /6) UHVXOWV JLYH ERQGEUHDNLQJ HQHUJLHV RI T9%% H9 DQG VWDQGDUG GHYLDWLRQ RI b IRU Q0267 DQG H9 DQG b IRU S0267 7KH +)+L QSQ%-7 GDWD LQYHUWHG WULDQJOH LQ )LJXUH f ZDV DOVR OHDVWVTXDUHILWWHG XVLQJ (TQf 7KH /6) UHVXOWV JLYHV D ERQGEUHDNLQJ HQHUJ\ RI T9%% H9 DQG D VWDQGDUG GHYLDWLRQ RI b $ UHDVRQDEOH EXLOWLQ SRWHQWLDO RI 9%, 9 ZDV DOVR REWDLQHG IURP WKH /6) UHIOHFWLQJ WKH KLJKO\ GRSHG EDVH UHJLRQ LQ DGYDQFHG VXEPLFURQ

PAGE 100

n+LWR )LJXUH &RPSDULVRQ RI H[SHULPHQWDO DQG WKHRUHWLFDO ERQGEUHDNLQJ HIILFLHQF\ E\ KRW HOHFWURQV DQG KROHV LQ 6L QSQ %-7 Q0267 DQG S0267 YHUVXV DSSOLHG MXQFWLRQ YROWDJH \HDU 77) LV WKH 77) DW WKH DSSOLHG MXQFWLRQ YROWDJH LQGLFDWHG E\ WKH YHUWLFDO DUURZ H[WUDSRODWHG E\ WKH WKHRU\ IURP WKH 77) GDWD DW WKH KLJKHU 95 IRU b RU b $,% $GDSWHG IURP 6DK > S @

PAGE 101

QSQ%-7 )RU WKH ,7+L QSQ%-7 GDWD (TQf ZDV XVHG LQVWHDG WR DFFRXQW IRU WKH YDOHQFHHOHFWURQ WXQQHOLQJ SURFHVV 7KH /6) UHVXOWV JLYHV D ERQGEUHDNLQJ HQHUJ\ RI H9 DQG D VWDQGDUG GHYLDWLRQ RI b $OVR LQFOXGHG LQ )LJXUH LV WKH WKHRUHWLFDO LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ IRU WKH /)(L SDWKZD\ ZLWK T9%% H9 DVVXPHG LQ WKH FDOFXODWLRQ $OEHLW WKH GLIIHUHQFH LQ Q0267 DQG S0267 IDEULFDWLRQ WHFKQRORJ\ WKH /6)fV ERQGEUHDNLQJ HQHUJLHV REWDLQHG IURP WKH Q0267 DQG S0267 GDWD DUH FRPSDUDEOH WR WKH WKHRUHWLFDO H9 6L+ ERQG HQHUJ\ >@ 7ZR LPSOLFDWLRQV FDQ EH PDGH IURP WKLV REVHUYDWLRQ f WKH RULJLQ RI WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDS LV WKH 6L+ ERQG DW WKH 6L6L LQWHUIDFH DQG f WKH PHFKDQLVP RI LQWHUIDFH WUDS JHQHUDWLRQ LV GXH WR WKH HQHUJHWLF KROHV EUHDNLQJ WKH ZHDN 6L+ ERQGV GXULQJ Q0267 &+( VWUHVV DQG QSQ %-7 UHYHUVH HPLWWHU VWUHVV DQG WKH HQHUJHWLF HOHFWURQ EUHDNLQJ WKH 6L+ ERQGV GXULQJ S0267 &++ VWUHVV 7KLV ERQGEUHDNLQJ LQWHUIDFH WUDS JHQHUDWLRQ SURFHVV FDQ EH GHVFULEHG E\ WKH IROORZLQJ NLQHWLF HTXDWLRQ 6L + Fr 6L +W F f ZKHUH Fr UHSUHVHQWV D HQHUJHWLF KRW FDUULHU KROH LQ Q0267 DQG QSQ %-7 DQG HOHFWURQ LQ S0267f +W LQGLFDWHV WKH GLIIXVLRQ RI K\GURJHQ DZD\ IURP WKH EURNHQ 6L ERQG 7KH FDUULHU F KDV WKHUPDO HQHUJ\ DIWHU WKH ERQGEUHDNLQJ SURFHVV )RU WKH QSQ%-7V WKH WZR +)+L DQG ,7+L SDWKZD\V \LHOG VLPLODU ERQG EUHDNLQJ HQHUJ\ T9%% H9f ZKLFK LV H[SHFWHG VLQFH WKH SULPDU\ FDUULHU UHVSRQVLEOH LQ QSQ%-7 GHJUDGDWLRQ LV WKH HQHUJHWLF KROH 7KH SULPDU\ GLIIHUHQFH LQ ERQG EUHDNLQJ HQHUJ\ EHWZHHQ WKH 0267 H9f DQG WKH QSQ%-7 H9f LV DWWULEXWHG WR WKH

PAGE 102

TXDOLW\ RI WKH 6L R[LGH SDVVLYDWLQJ WKH 6L6L LQWHUIDFH )RU 0267 WKLV R[LGH LV WKHUPDOO\ JURZQ RQ WKH 6L VXEVWUDWH ZKLFK LV LQKHUHQWO\ EHWWHU WKDQ WKH FKHPLFDOO\ GHSRVLWHG R[LGH XVHG WR FRYHU WKH HPLWWHUEDVH LQWHUIDFH RI WKH QSQ%-7 'LIIHUHQFH LQ WKH VXUIDFH LPSXULW\ FRQFHQWUDWLRQ SURILOH DVVRFLDWHG ZLWK WKH IDEULFDWLRQ RI WKH QSQ%-7 DQG WKH 0267 PD\ DOVR DIIHFW WKH ERQGEUHDNLQJ HQHUJ\

PAGE 103

&+$37(5 3+<6,&6%$6(' 7,0(72)$,/85( (;75$32/$7,21 $/*25,7+0 86,1* &855(17$&&(/(5$7(' &+$11(/ +27 &$55,(5 675(66 ,QWURGXFWLRQ 9ROWDJHDFFHOHUDWLRQ GXULQJ FKDQQHO KRW FDUULHU &+&f VWUHVV KDV EHHQ HPSOR\HG LQGXVWU\ZLGH WR GHWHUPLQH WKH PD[LPXP RSHUDWLRQ YROWDJH SHUPLWWHG IRU D PHWDOR[LGH VLOLFRQ WUDQVLVWRU 0267f WR PHHW WKH RSHUDWLRQ WHQ\HDU WLPHWRIDLOXUH 77)3f UHTXLUHPHQW >@ ,Q WKLV PHWKRG WKH IDLOXUH UDWH RI 77) GDWD REWDLQHG DW KLJKHU VWUHVV YROWDJHV DQG VKRUW VWUHVV WLPHV DUH H[WUDSRODWHG WR GHWHUPLQH WKH 77)3 DW WKH GHVLUHG ORZHU RSHUDWLRQ YROWDJH ,W LV EDVHG RQ WKH DVVXPSWLRQ WKDW GHJUDGDWLRQ PHFKDQLVP DW WKH KLJKHU VWUHVV YROWDJHV DSSOLHG WR 0267fV GUDLQERG\ MXQFWLRQ DQG ELSRODU MXQFWLRQ WUDQVLVWRUfV %-7fVf HPLWWHUEDVH MXQFWLRQ DUH LQGHSHQGHQW RI WKH PDJQLWXGH RI WKH VWUHVV YROWDJHV 7UDQVLVWRUVf GHJUDGDWLRQ KDV EHHQ DWWULEXWHG WR WKH JHQHUDWLRQ RI R[LGHVLOLFRQ LQWHUIDFH WUDSV DQG FKDUJLQJ t JHQHUDWLRQ RI R[LGH WUDSV E\ WKH HQHUJHWLF RU KRW HOHFWURQV DQG KROHV >@ 7KH YROWDJHH[WUDSRODWLRQ IURP KLJKVWUHVV YROWDJH WDFLWO\ DVVXPHV WKDW WKH IXQGDPHQWDO IDLOXUH PHFKDQLVPV GXH WR KRW FDUULHUV DUH LQGHSHQGHQW RI WKH DSSOLHG YROWDJHV ZKLFK LV XQWHQDEOH EHFDXVH WKH GHJUDGDWLRQ PHFKDQLVPV KDYH KRWFDUULHU NLQHWLF HQHUJ\ RU DSSOLHG YROWDJH WKUHVKROG > @ EHORZ ZKLFK WKH GHJUDGDWLRQ PHFKDQLVPV FHDVH 7R FKDUDFWHUL]H WKH GHJUDGDWLRQ PHFKDQLVPV RI %-7 WHFKQRORJLHV DW WKH RSHUDWLRQ YROWDJH 9f ZLWK

PAGE 104

UHODWLYHO\ VKRUW WXUQRYHU WLPH 1HXJURVFKHO DQG 6DK SURSRVHG WZR PHWKRGV WR LQFUHDVH WKH VWUHVV FXUUHQW >@ DW ORZ VWUHVV YROWDJH DV FRQWUDVW WR KLJK YROWDJH VWUHVV WR DFFHOHUDWH WKH GHJUDGDWLRQ UDWH RI WKH WUDQVLVWRUV DQG XVLQJ WKH '&,9 PHDVXUHPHQW WR PRQLWRU WKH KRW FDUULHULQGXFHG GHJUDGDWLRQ ,Q WKLV FKDSWHU ZH ZLOO GHPRQVWUDWH WKDW WKH %-7 FXUUHQWDFFHOHUDWLRQ PHWKRGV FDQ EH H[WHQGHG WR 0267V XQGHU &+& VWUHVV )LUVW ORQJ FKDQQHO 0267Vf VWUXFWXUH ZLOO EH XVHG IRU GHPRQVWUDWLQJ WKH FXUUHQW DFFHOHUDWHG FKDQQHO KRW FDUULHU VWUHVV PHWKRGRORJ\ /RQJ FKDQQHO 0267V DUH XVHG EHFDXVH WKH WUDQVLVWRU LV HVVHQWLDOO\ DQG WKLV ZLOO VLPSOLI\ DQG DLG LQ WKH DQDO\VLV 8SRQ OD\LQJ RXW WKH IUDPHZRUN DQG XQGHUVWDQGLQJ WKH EDVLV SK\VLF RI WKH FXUUHQW DFFHOHUDWHG FKDQQHO KRW FDUULHU VWUHVV &$&+&f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f GHQVLW\ DQG f NLQHWLF HQHUJ\ RI KRW FDUULHUV f WKH WKUHVKROG NLQHWLF HQHUJ\ RI GHJUDGDWLRQ SDWKZD\ DQG WKH GHQVLW\ RI UHVLGXDOIDEULFDWLRQ FUHDWHG f DFWLYH RU QRQK\GURJHQDWHG DQG f SDVVLYH RU K\GURJHQDWHG 6L+ DQG 6L2+f R[LGH DQG LQWHUIDFLDO HOHFWURQ DQG KROH WUDSV 7UDQVLVWRU GHJUDGDWLRQ UDWH PD\ EH LQFUHDVHG E\ LQFUHDVLQJ WKH KRW FDUULHU NLQHWLF HQHUJ\ XVLQJ KLJKHU VWUHVV YROWDJH WKH WUDGLWLRQDO YROWDJH DFFHOHUDWLRQ PHWKRGf RU ODUJHU KRW FDUULHU

PAGE 105

FXUUHQW GHQVLW\ WKH FXUUHQWDFFHOHUDWLRQ PHWKRGf ,Q WKH FXUUHQWDFFHOHUDWLRQ PHWKRG WKH KRW FDUULHU GHQVLW\ FDQ EH LQFUHDVHG E\ IRUZDUGLQJELDVLQJ D SQ MXQFWLRQ ZKLOH WKH KRW FDUULHU NLQHWLF HQHUJ\ LV FRQWUROOHG E\ WKH DSSOLHG YROWDJHV GXULQJ VWUHVV 8QGHU WKH WUDGLWLRQDO FKDQQHO KRW HOHFWURQ &+(f VWUHVV RI D QFKDQQHO 0267 Q0267f WKH KLJK GUDLQ DQG ORZ JDWH YROWDJHV 9'!9*f DFFHOHUDWH WKH VRXUFHLQMHFWHG FKDQQHO HOHFWURQV GHQRWHG E\ ,F+VRXUFH rQ )LJXUH WR KLJK NLQHWLF HQHUJLHV TO9' Y* 9*7f, LQ WKH VSDFHFKDUJHUHJLRQ RI WKH UHYHUVHELDVHG GUDLQ MXQFWLRQ $V VKRZQ LQ )LJXUH WKHVH KRW HOHFWURQV FDQ JHQHUDWH LQWHUIDFH WUDSV E\ EUHDNLQJ WKH 6L+ DQG 6L2+ ERQGV >@ WKH VWDUILOOHG WULDQJOH r LQ )LJXUH f WR UHOHDVH WKH K\GURJHQ DQG OHDYH WKH GDQJOLQJ VLOLFRQ ERQGV ZKLFK DUH HOHFWURQ DQG KROH LQWHUIDFH WUDSV 7KH FKDQQHO KRW HOHFWURQV ZLOO DOVR JHQHUDWH DGGLWLRQDO VHFRQGDU\f HOHFWURQKROH SDLUV .(!(J H9 DV VKRZQ LQ >@f QHDU WKH GUDLQ MXQFWLRQ 7KH VHFRQGDU\ KRW KROHV FDQ EUHDN DGGLWLRQDO 6L+ DQG 6L2+ ERQGV WR JLYH PRUH LQWHUIDFH WUDSV >@ ,I WKH VHFRQGDU\ KRW KROHV KDYH .( M!;K H9 WKH 6L6L KROH EDUULHUf WKH\ FDQ DOVR EH LQMHFWHG LQWR WKH R[LGH WR SRVLWLYHO\ FKDUJH WKH H[LVWLQJ R[LGH KROH WUDSV )LJXUH LOOXVWUDWHV WKH FXUUHQW DFFHOHUDWLRQ PHWKRG ,W VKRZV WKDW WKH FKDQQHO KRW HOHFWURQ FXUUHQW ,&+HLHFWURQ IF+VRXUFH O&+EH! FDQ EH LQFUHDVHG LQ WZR ZD\V f LQFUHDVLQJ ,F+VRXUFH E\ IRUZDUGELDVLQJ WKH VRXUFHERG\ RU 6RXUFH(PLWWHU 6(f QS MXQFWLRQ MXQFWLRQ 9%6 9f DQG f LQFUHDVLQJ ,*+EH E\ IRUZDUGELDVLQJ WKH VXEVWUDWHERG\ RU %RWWRP(PLWWHU %(f QS MXQFWLRQ 9%( 9 RU 9f )LJXUH GHPRQVWUDWHV LQFUHDVHG FKDQQHO KRW HOHFWURQ FXUUHQW IURP IRUZDUGELDVLQJ WKH ERWWRPHPLWWHU MXQFWLRQ ,F+HLHFWURQ ,F+EHff E\ D IDFWRU RI DERXW DW 9%( 9 DQG

PAGE 106

96% 229 RU 9 6L2" *DWH f 9JV 9 'UDLQ 96 9 WR 9 !' %DVH LWUA *URXQG 6666 6L2] 1LU r 6Lrr+ $ $ $ B $ $ $ $ $ $ c666 $ Q f§ &+VRXUFH &+EH A&+HOHFWURQ Q $ S%DVHZHOO RU S%RG\ n-' %XON '%XON $ t &76 f Q%RWWRP(PLWWHU (PLWWHU A 9(% 9 RU 9 RU 9 )LJXUH &URVVVHFWLRQDO YLHZ RI D Q0267 ELDVHG VLPXOWDQHRXVO\ LQ WKH WZR FXUUHQW DFFHOHUDWHG &+( VWUHVV FRQILJXUDWLRQV 'RWV DUH FRQGXFWLRQEDQG HOHFWURQV &LUFOHG GRWV DUH YDOHQFHEDQG HOHFWURQV &LUFOHV DUH YDOHQFHEDQG KROHV JHQHUDWHG E\ HQHUJHWLF HOHFWURQ LPSDFW LQ WKH Q GUDLQ >@ +ROORZ WULDQJOHV DUH LQWHUIDFH HOHFWURQLF WUDSV 6ROLG WULDQJOHV DUH ZHDN LQWHUIDFH ERQGV ZKLFK EHFRPH HOHFWURQLF LQWHUIDFH WUDSV ZKHQ UXSWXUHG E\ KROH FDSWXUH PDUNHG E\ r ZKLFK UHOHDVHV WKH ERQGHG K\GURJHQ 6ROLG DUURZ ZLWK D GRW RQ LWV WDLO GHVLJQDWHV DQ HOHFWURQ LQMHFWHG IURP WKH IRUZDUGELDVHG ERWWRP HPLWWHU RU VRXUFH MXQFWLRQV +ROORZ DUURZ ZLWK D FLUFOH RQ LWV WDLO GHVLJQDWHV WKH KROH SDWKZD\ 5HJXODU FKDQQHO HOHFWURQV DQG WKHUPDO HOHFWURQV DUH QRW VKRZQ %LDVHV DUH WKRVH XVHG WR JLYH WKH GDWD LQ )LJV

PAGE 107

&+HOHFWURQ $f )LJXUH (IIHFW RI IRUZDUG DSSOLHG YROWDJH WR WKH ERWWRPHPLWWHU MXQFWLRQ 9%( RQ WKH FKDQQHO KRW HOHFWURQ FXUUHQW YHUVXV GUDLQ YROWDJH FKDUDFWHULVWLFV A&+HOHFWURQnA'6f DW 9*6 A 9 DQG 96%

PAGE 108

DW 9EH 9 ,F+HOHFWURQ ZDV REWDLQHG E\ VXEWUDFWLQJ WKH EXON DQG H[WHULRU SHULPHWHU HOHFWURQ FXUUHQWV IURP WKH WRWDO GUDLQ WHUPLQDO FXUUHQW ,' PHDVXUHG DW IRXU FRPELQDWLRQV RI HPLWWHU DQG JDWH YROWDJHV IRUZDUG DQG ]HUR ELDV WKH VXEVWUDWHERG\ MXQFWLRQ ZLWK DQG ZLWKRXW WKH LQMHFWHG HOHFWURQV IURP WKH ERWWRP HPLWWHUf DQG LQYHUVLRQ DQG DFFXPXODWLRQ JDWH YROWDJHV ZLWK DQG ZLWKRXW WKH FKDQQHO HOHFWURQV DW WKH 6L6L LQWHUIDFHf 5HVXOWV DQG 'LVFXVVLRQV /RQJ &KDQQHO &$&+& 3URGXFWLRQ Q0267V DQG S0267V VXSSOLHG E\ VHYHUDO 65& FRPSDQLHV DUH XVHG WR GHPRQVWUDWH WKH SURSRVHG FXUUHQWDFFHOHUDWHG FKDQQHO KRW HOHFWURQ &$&+(f DQG FKDQQHO KRW KROH &$&++f VWUHVV PHWKRGRORJLHV 7KH Q0267 RQ SZHOOQ EXULHG OD\HU KDV QP JDWH R[LGH :/ LP _LP DQG 1$$BZHOO [O %RURQFP 7KH S0267 RQ QZHOOSVXEVWUDWH KDV QP JDWH R[LGH :/ _LP ILP DQG 1''ZHLL [O 3KRVSKRUXVFP ,Q WKH 6$0 H[SHULPHQWV WKH GHJUDGDWLRQ ZDV PRQLWRUHG E\ WKH 'UDLQ(PLWWHU '&,9 '('&,9f PHWKRG ZKLFK PHDVXUHV WKH SRVWn VWUHVV EDVH RU ERG\ FXUUHQW YHUVXV JDWHEDVH YROWDJH FKDUDFWHULVWLFV ,%9*% ZLWK WKH GUDLQERG\ MXQFWLRQ IRUZDUG ELDVHG )LJXUH VKRZV WKH PHDVXUHG 77)fV GHILQHG EHORZf YHUVXV DSSOLHG GUDLQVRXUFH VWUHVV YROWDJH 9'6 ERWK ZLWKRXW 9%( 9%6 9f DQG ZLWK 9%(! RU 9%6!f FXUUHQW DFFHOHUDWLRQ 7KH SRVWVWUHVV '&,9 FXUYHV DUH VKRZQ LQ WKH LQVHW RI )LJXUH Df IRU Q0267 DQG )LJXUH Ef IRU S0267 7KH PHDVXUHG 77) LV GHILQHG DV WKH VWUHVV WLPH UHTXLUHG WR JLYH OHSHDNA%LQLWLDUA RU D b LQFUHDVH LQ ,% IRU Q0267 DQG

PAGE 109

)LJXUH Df 7LPHWR)DLOXUH YHUVXV GUDLQ VWUHVV YROWDJH RI LP[ _LP Q0267 IRU QRFXUUHQWDFFHOHUDWLRQ VWUHVV IRXU FLUFOHVf DQG FXUUHQWDFFHOHUDWHG VWUHVV ILYH VROLG GRWV DW 9%( 9 RQH VROLG WULDQJOH DW 9%( 9 DQG DQ DVWHULVNV DW 96% 9 $OO WUDQVLVWRUV DUH VWUHVVHG DW 9&6 9 7KH LQVHW VKRZV WKH '('&,9 ,En9JEf FKDUDFWHULVWLFV PHDVXUHG DW 9GE 9 9VE 9HE 9f DV D IXQFWLRQ RI &+( VWUHVV WLPH DQG VHFf 6WUHVVHG DW 9'6 9 9*6 9 DQG 96% 9%( 9

PAGE 110

V2GOL 9'69f )LJXUH Ef 7LPHWR)DLOXUH YHUVXV GUDLQ VWUHVV YROWDJH RI MLP[_LP S0267 IRU QRFXUUHQWDFFHOHUDWLRQ VWUHVV IRXU FLUFOHVf DQG FXUUHQWDFFHOHUDWHG VWUHVV ILYH VROLG GRWV DW 9(% 9f $OO WUDQVLVWRUV DUH VWUHVVHG DW 9*6 9 7KH LQVHW VKRZV WKH '('&,9 ,%9*%f FKDUDFWHULVWLFV PHDVXUHG DW 9GE 9 96% 9(% 9f DV D IXQFWLRQ RI &++ VWUHVV WLPH DQG [O VHFf 6WUHVVHG DW 9'6 9 9*6 9 DQG 96% 9%( 9

PAGE 111

A%SHDN$%LQLWLDL RU D b LQFUHDVH LQ ,% IRU S0267 7KH JURZWK RI WKH ,%SHDN ZLWK LQFUHDVLQJ VWUHVV WLPH LV DWWULEXWHG WR LQWHUIDFH WUDS JHQHUDWLRQ > @ E\ KRW FDUULHUV KRW KROHV IRU Q0267 DQG KRW HOHFWURQV IRU S0267f EUHDNLQJ WKH 6L+ DQG 6L2+ ERQGV QHDU WKH GUDLQ MXQFWLRQ >@ 7KH SURILOHV RI WKH ,% SHDN DW 9*% 9 IRU Q0267 DQG 9*% 9 IRU S0267 UHPDLQ QHDUO\FRQVWDQW LQGLFDWLQJ WKDW QHJOLJLEOH LQWHUIDFH WUDSV ZHUH EHLQJ JHQHUDWHG RYHU WKH FKDQQHO 7KH 9*6VKLIW RI WKH c%SHDN DWWULEXWHG WR FKDUJLQJ RI WKH R[LGH KROH WUDSV QHJDWLYH9* VKLIW LQ Q0267f DQG HOHFWURQ WUDSV SRVLWLYH9* VKLIW LQ S0267f IURP EDUULHUVXUPRXQWLQJ LQMHFWLRQ LQWR WKH R[LGH RI WKH VHFRQGDU\ KRW FDUULHUV LQWHUEDQGLPSDFWJHQHUDWHG E\ WKH SULPDU\f FKDQQHO KRW FDUULHUV ZKLFK JDYH WKH SUHGLFWHG WKUHVKROG H9 IRU KROHV DQG H9 IRU HOHFWURQV 'XULQJ WKH &$&+& VWUHVV XVLQJ WKH IRUZDUGELDVHG ERWWRP HPLWWHU MXQFWLRQ D FRQVWDQW LQFUHDVH LQ WKH EDVHOLQH RI WKH ,% FXUUHQW LV REVHUYHG 7KLV LV DWWULEXWHG WR WKH LQWHUIDFH WUDS JHQHUDWHG DORQJ WKH H[WHULRU WR WKH FKDQQHOf SHULSKHUDO RI GUDLQ MXQFWLRQ XQGHUQHDWK WKH ILHOG R[LGH 7KLV H[FHVV ,% FRPSRQHQW FUHDWHG GXULQJ &$&+& VWUHVV DUH VXEWUDFWHG LQ RUGHU WR DFFRXQW IRU WKH DFWXDO &+(LQGXFHG GHJUDGDWLRQ RFFXUULQJ DW WKH GUDLQ MXQFWLRQ LQWHULRU WR WKH FKDQQHO )LJXUH Df FRQWDLQV WKUHH VHWV RI 77) GDWD f WKH WUDGLWLRQDO &+( VWUHVV ZLWKRXW FXUUHQWDFFHOHUDWLRQ &$f IRXU FLUFOHV IRU 9'6BVWUHVV 9 WR 9f f &$&+( VWUHVV IURP IRUZDUGELDVLQJ WKH ERWWRP HPLWWHU VXEVWUDWHERG\ QS MXQFWLRQ ILYH VROLG FLUFOHV DW 9%( 9 IRU 9'6BVWUHVV 9 WR 9 JLYLQJ DERXW [ DFFHOHUDWLRQ DQG RQH VROLG WULDQJOH DW 9%( 9 IRU 9'6 9 JLYLQJ [ DFFHOHUDWLRQf DQG f &$&+( VWUHVV IURP IRUZDUGELDVLQJ WKH WRSHPLWWHU VRXUFHERG\

PAGE 112

QS MXQFWLRQ RQH DVWHULVNV DW 9%6 9 IRU 9'6 9 JLYLQJ DERXW [ DFFHOHUDWLRQf $W HDFK DSSOLHG 9'6BVWUHVV WKH DSSOLHG 9*6VWUHVV LQ f DQG f ZDV NHSW WKH VDPH DV LQ f 5HGXFWLRQ LQ 6$0 WLPHV WR PHDVXUH 77) XVLQJ &$&+( VFDOHV ZLWK WKH LQFUHDVHG &+( FXUUHQW E\ IRUZDUGELDVLQJ WKH VXEVWUDWHERG\ RU VRXUFHERG\ QS MXQFWLRQ )RU H[DPSOH DW 9'6VWUHVV 9 DQG 9%( 9 77)QR&$&+(77)&+( VV DQG ,&+HOHFWURQ9%( 9f,&+HHFWURQ9%( 9f IO$ S$ :LWKRXW FXUUHQWDFFHOHUDWLRQ WKH HVWLPDWHG 6$0 WLPH DW 9'6VWUHVV 9 LV [V GD\V XQILOOHG WULDQJOHf IURP H[WUDSRODWLQJ WKH IRXU QR&$&+( 77) GDWD SRLQWV FLUFOHVf WR 9 &XUUHQW DFFHOHUDWLRQ YLD 9%( 9 JDYH D [ UHGXFWLRQ RI WKH VWUHVV WHVW WLPH WR [O V GD\V ILOOHG WULDQJOHf 7KXV WKLV FXUUHQWDFFHOHUDWLRQ H[SHULPHQW KDV GHPRQVWUDWHG WKH SUDFWLFDO DSSOLFDWLRQ LQ UHGXFLQJ WKH PHDVXUHPHQW WLPH IURP GD\V RU \HDUV WR GD\V RU ZHHNV LQ RUGHU WR GHWHUPLQH WKH \HDU RSHUDWLRQ 77) RI WKLV WUDQVLVWRU WHFKQRORJ\ 7KH &$&++ 77) GDWD RI WKH S0267 LQ )LJXUH Ef ZHUH REWDLQHG E\ IRUZDUGn ELDVLQJ WKH VXEVWUDWHERG\ SQ MXQFWLRQ WR 9%( 9 ZKLFK VKRZHG RQO\ D O2[ UHGXFWLRQ 7KLV LV VPDOOHU WKDQ WKH DFFHOHUDWLRQ REVHUYHG LQ Q0267 LQ )LJXUH Df DW D ORZHU IRUZDUG ELDVHV 9%( 9 [f DQG 9 [f DQG LW LV DWWULEXWHG WR WKH ODUJHU VHULHV EDVH UHVLVWDQFH RI WKH QZHOO DQG SVXEVWUDWH ZKLFK UHGXFHG WKH IRUZDUG FXUUHQW 7KH 77)TSfV LQ )LJXUH DW WKH GHVLJQHG ORZHU RSHUDWLRQ YROWDJHV ZLWK QR FXUUHQW DFFHOHUDWLRQ ZHUH REWDLQHG IURP WKH DFFHOHUDWHG DQG QRWDFFHOHUDWHG IOXHQFHWR

PAGE 113

U7)23Vf ,9'6,9f )LJXUH )7) )OXHQFHWR)DLOXUHf RQ ULJKW \D[LV DQG 77)3 RSHUDWLRQ77) LH ZLWKRXW FXUUHQWDFFHOHUDWLRQf RQ OHIW \D[LV YHUVXV WKH PDJQLWXGHV RI 9'6 6ROLG FXUYHV DUH WKH /6) RI WKH Q0267 DQG S0267 GDWD WR WKH WKHRU\ JLYLQJ ERQGEUHDNLQJ NLQHWLF HQHUJLHV RI T9%% H9 DQG H9 UHVSHFWLYHO\ )7)TFPf

PAGE 114

IDLOXUH )7) ODEHOHG RQ WKH ULJKW YHUWLFDO D[LVf $W HDFK VWUHVV LQWHUYDO WKH )7) LV FRPSXWHG XVLQJ )7) T -M f ZKHUH -F+HLHFWURQ A&+HLHFWURQAGUDLQ DQFr WKH GUDLQ FRQWDFW DUHD ZDV XVHG 7KH RSHUDWLRQ 77) RQ WKH OHIW \D[LVf LV WKHQ REWDLQHG IURP 77)S T)7)9(%f-&+BHOHFWURQ9(% f f 7KH YDOLGLW\ RI WKH FXUUHQW DFFHOHUDWLRQ PHWKRGRORJ\ LV FRQILUPHG IRU ERWK Q0267 DQG S0267 E\ WKH GDWD RYHUODS FRYHULQJ PDQ\ GHFDGHV RI WKH QRQDFFHOHUDWHG 77)V RSHQ FLUFOHVf ZLWK WKH DFFHOHUDWHG 77)V DVWHULVN VROLG WULDQJOH DQG FLUFOHVf LQ WKH KLJK YROWDJH UDQJH 9 A,9'6VWUHVVO 9f ,W DOVR VXSSRUW WKH DQWLFLSDWLRQ WKDW D VPDOO LQFUHDVH LQ FDUULHU GHQVLW\ DQG FXUUHQW LQ WKH FKDQQHO IURP D IRUZDUGELDVHG MXQFWLRQ ZRXOG QRW VLJQLILFDQWO\ DOWHU WKH VSDWLDO YDULDWLRQ RI WKH HOHFWULF ILHOG KRW HOHFWURQ DQG KROH GHQVLWLHV DQG WKHLU NLQHWLF HQHUJLHV 7KH WZR VROLG FXUYHV LQ )LJXUH ZHUH REWDLQHG IURP D OHDVWVTXDUHILW /6)f RI WKH HOHYHQ Q0267 GDWD SRLQWV DQG QLQH S0267 GDWD SRLQWV WR WKH WKHRU\ SURSRVHG E\ 6DK (TQ ff >@ WR DFFRXQW IRU LQWHUIDFH WUDS JHQHUDWLRQ GXULQJ FKDQQHO KRW FDUULHU VWUHVV 8VLQJ D IDLOXUH FULWHULD RI $,%,% b IRU Q0267 DQG $,%,% b IRU S0267 WKH WKHRUHWLFDOO\ H[WUDSRODWHG PDJQLWXGH RI WKH PD[LPXP 9'6 YROWDJHV WR JLYH 77)S \HDUV DUH 9 DQG 9 UHVSHFWLYHO\ IRU WKHVH WZR WHFKQRORJLHV $Q LPSRUWDQW FRQILUPDWLRQ RI WKH FXUUHQWDFFHOHUDWLRQ PHWKRGRORJ\ DQG DOVR WKH ERQGn EUHDNLQJ WKHRU\ LV WKDW WKH /6)fV RI WKH Q0267 DQG S0267 WR WKHRU\ SUHVHQWHG LQ &KDSWHU JDYH QHDUO\ LGHQWLFDO K\GURJHQERQGEUHDNLQJ WKUHVKROG HQHUJ\ 9%%

PAGE 115

H9 M bf E\ KRW KROHV LQ Q0267 DQG H9 D bf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

PAGE 116

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fV RI WKH ,6VDW DQG $9*W YHUVXV GUDLQVRXUFH VWUHVV YROWDJHV 9'6 LV VKRZQ LQ )LJXUH Df 7KH IDLOXUH FULWHULD XVHG WR GHWHUPLQH WKH 77)fV DUH GHILQHG DV D b UHGXFWLRQ LQ ,'VDW DQG D P9 LQFUHDVH LQ $*7 $ [ VWUHVVWLPH UHGXFWLRQ LV REVHUYHG IRU WKH Q0267 VWUHVVHG XQGHU &$&+( FRQGLWLRQ ,Q )LJXUH Ef WKH )7) GDWD DUH ILWWHG WR 6DKfV LQWHUIDFH WUDS WKHRU\ XVLQJ (TQ f ZKHUH WKH ERQGEUHDNLQJ HQHUJ\ T9%% LQ (TQ f LV UHSODFHG E\ DQ HIIHFWLYH ERQGEUHDNLQJ HQHUJ\ T9%% WR WDNH LQWR DFFRXQW RI $XJHUKRW KROH LQMHFWLRQ IURP WKH GUDLQ HGJH LQWR WKH GUDLQ VSDFHFKDUJHUHJLRQ WR JHQHUDWH LQWHUIDFH WUDS $V ZLOO EH VKRZQ LQ &KDSWHU KRW KROH JHQHUDWHG E\ $XJHU UHFRPELQDWLRQ RI FKDQQHO KRW HOHFWURQ DQG WKHUPDO KROH LQ WKH QGUDLQ TXDVLQHXWUDO UHJLRQ UHTXLUHV WKH ORZHVW HOHFWURQ NLQHWLF HQHUJ\ +HQFH 9%% LV UHODWHG WR 9%% E\ A%%n a A%% A6L*DSA f ZKHUH 95 9' 9'VDW DQG (6LJDS LV WKH VLOLFRQ HQHUJ\ JDS 7KH LQWHUIDFH WUDS JHQHUDWLRQ HIILFLHQF\ GXH WR KRW KROH IURP $XJHU UHFRPELQDWLRQ SDWKZD\ LV WKHQ JLYHQ E\

PAGE 117

5B9%%(VLJDST@ f 7KH RSHUDWLRQ 77) FRPSXWHG XVLQJ (TQ f LV VKRZQ RQ WKH OHIWD[LV RI )LJXUH Ef 8VLQJ WKH IDLOXUH FULWHULD RI b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f f ZKHUH $ DQG % DUH HPSLULFDO ILWWLQJ FRQVWDQWV $ FORVHU H[DPLQDWLRQ RI (TQ f JLYHV D PRVW LPSRUWDQW UHVXOW GHULYHG E\ 6DKf >@ 7KH KROH LQMHFWLRQ IURP WKH GUDLQ HGJH GXH WR WKLV $XJHUUHFRPELQDWLRQ SDWKZD\ >@ KDV WKH ORZHVW WKUHVKROG IRU LQWHUIDFH WUDS JHQHUDWLRQ 7KLV LPSOLHV WKDW LQWHUIDFH WUDS JHQHUDWLRQ ZLOO UHPDLQ D GRPLQDQW GHJUDGDWLRQ SURFHVV IRU ORZYROWDJH 0267 DV ORQJ DV WKH VXSSO\ YROWDJH LV JUHDWHU WKDQ ,9 7KLV LV LOOXVWUDWHG E\ WKH IROORZLQJ UHODWLRQ 9' 9'VDW 9%% f f ,9

PAGE 118

:KHQ DSSOLHG WR WKH %-7 XQGHU HPLWWHUEDVH UHYHUVHELDV VWUHVV LQ ZKLFK 9%, LV QR ORQJHU ]HUR WKH WKUHVKROG GURSV WR ]HUR DV VKRZQ E\ 6DK >@

PAGE 119

,OO 96 9* 9f )LJXUH 'HJUDGDWLRQ FKDUDFWHULVWLFV RI _LP Q0267 XQGHU &+( VWUHVV ,'VDW LV PHDVXUHG LQ WKH UHYHUVH LH DFWXDOO\ ,6VDW ZDV PHDVXUHGf FRQILJXUDWLRQ WR PRQLWRU WKH GHJUDGDWLRQ QHDU WKH GUDLQ MXQFWLRQ &+( VWUHVV FRQGLWLRQ 9'6 9 9*6 9 DQG 96% 9

PAGE 120

94' 9f )LJXUH 'HJUDGDWLRQ FKDUDFWHULVWLFV RI ILP Q0267 XQGHU &+( VWUHVV ,'9* LV PHDVXUHG LQ WKH UHYHUVH FRQILJXUDWLRQ WR PRQLWRU WKH GHJUDGDWLRQ QHDU WKH GUDLQ MXQFWLRQ 0HDVXUHG DW 9'6 9 &+( VWUHVV FRQGLWLRQ 9'6 9 9*6 9 96% 9

PAGE 121

)LJXUH '('&,9 'HJUDGDWLRQ FKDUDFWHULVWLFV RI _LP Q0267 XQGHU &+( VWUHVV &+( VWUHVV FRQGLWLRQ 9'6 9 9*6 9 96% 9

PAGE 122

'VDU 'VDW2 r r &7 &7 r b$? %'(SHDN )LJXUH 'HSHQGHQFH RI $,'BVDW,'BVDW $9*; RQ $,%SHDN,%SHDN IRU SP Q0267 XQGHU &+( RSHQ DQG ILOOHG WULDQJOHVf DQG &$&+( RSHQ DQG ILOOHG FLUFOHVf &+( VWUHVV FRQGLWLRQ 9'6 9 9*6 9 96% 9 &$&+( VWUHVV nFRQGLWLRQ 9'6 9 9*6 9 96% 9 $9JW9f

PAGE 123

9'69f )LJXUH Df 7LPHWR)DLOXUH YHUVXV GUDLQ VWUHVV YROWDJH RI LP[>MP Q0267 IRU QRFXUUHQWDFFHOHUDWLRQ VWUHVV RSHQ WULDQJOHV DQG FLUFOHVf DQG FXUUHQW DFFHOHUDWHG VWUHVV VROLG WULDQJOHV DQG FLUFOHVf $OO WUDQVLVWRUV DUH VWUHVVHG DW 9JV 9 ,'VDW LV PHDVXUHG DW 9' 9* 9 DQG 9*7 LV H[WUDFWHG DW ,' P$ Ef )7) )OXHQFHWR)DLOXUHf RQ ULJKW \D[LV DQG 77)23 RSHUDWLRQ77) LH ZLWKRXW FXUUHQWDFFHOHUDWLRQf RQ OHIW \D[LV YHUVXV WKH PDJQLWXGHV RI 9'6 6ROLG FXUYH LV WKH /6) RI WKH Q0267 ,'VDW GDWD JLYLQJ ERQGEUHDNLQJ NLQHWLF HQHUJ\ RI T9%% H9 )7)TFPf

PAGE 124

R7 &r s LY'6L Yf )LJXUH 7KH WUDGLWLRQDO YROWDJHDFFHOHUDWHG 77) H[WUDSRODWLRQ SORW IRU WKH 9 _LP Q0267 WHFKQRORJ\

PAGE 125

&+$37(5 +27 +2/( ,1-(&7,21 ,172 6L ,1 1&+$11(/ 026 75$16,6725 '85,1* &+$11(/+27(/(&7521 675(66 ,QWURGXFWLRQ $ GHFUHDVH LQ WKUHVKROG YROWDJH > @ DQG DQ LQFUHDVH LQ WUDQVFRQGXFWDQFH >@ ZHUH REVHUYHG ZKHQ VKRUW QFKDQQHO PHWDOR[LGHVLOLFRQ WUDQVLVWRUV 0267f DUH VWUHVVHG XQGHU KLJK 9' DQG ORZ 9* a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f 7KH\ FRQFOXGHG WKDW DW KLJK9' DQG ORZ9* a9*7f VWUHVV FRQGLWLRQ WKH XOWUDORZ SRVLWLYH JDWH FXUUHQW PHDVXUHG LQ WKHLU H[SHULPHQW LV GXH WR KRWKROH VXUPRXQWLQJ WKH 6L6L HQHUJ\ EDUULHU _f+BK H9f DQG LQMHFWHG LQWR WKH 6L ,QMHFWHG KROHV FDQ EH

PAGE 126

FDSWXUHG E\ WKH R[LGH KROH WUDSV ZKLFK EHFRPHV SRVLWLYHO\ FKDUJHG >@ ,Q DGGLWLRQ FKDUJH SXPSLQJ PHDVXUHPHQWV DOVR VXJJHVW WKDW ORFDOL]HG KROH WUDSSLQJ LQWR WKH JDWH R[LGH FDQ FDXVH D QHJDWLYH YROWDJH VKLIW RI FKDUJHSXPSLQJ FKDUDFWHULVWLFV > @ $ VXUYH\ RI WKH ELDVLQJ FRQGLWLRQV UHSRUWHG LQ > @ VKRZHG KLJK GUDLQ YROWDJH 9 9' 9f DQG ORZ JDWH YROWDJH 9 9* 9f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f WKUHVKROG NLQHWLF HQHUJLHV IRU JHQHUDWLQJ D H9 KRW KROH

PAGE 127

VHFRQGDU\f ([SHULPHQW ZLOO VKRZ WKDW GXULQJ FKDQQHO KRW HOHFWURQ VWUHVV D H9 LQWHUEDQG LPSDFWJHQHUDWHG VHFRQGDU\ KRW KROH LV FUHDWHG ZKHQ D SULPDU\ LPSDFWLQJ HOHFWURQ SRVVHVVHV NLQHWLF HQHUJ\ RI PRUH WKDQ H9 )RXU )XQGDPHQWDO ,QWHUEDQG +RW +ROH *HQHUDWLRQ 3DWKZD\V )LJXUH VKRZV IRXU IXQGDPHQWDO LQWHUEDQG JHQHUDWLRQ SDWKZD\V WR FUHDWH D H9 KRW KROH ZLWK WKH KRW HOHFWURQ DV WKH LQLWLDWLQJ FDUULHU )RU HDFK SDWKZD\ VKRZQ LQ )LJXUH WKH HQHUJ\ RI WKH KRW HOHFWURQ DQG KROHV LV GUDZQ WRVFDOH 7KLV V\VWHPDWLF FODVVLILFDWLRQ RI JHQHUDWLQJ H9 KRW KROH E\ WKH SULPDU\ KRW HOHFWURQ ZDV GHOLQHDWHG E\ 6DK EDVHG RQ WKH IXQGDPHQWDO LPSDFW DQG $XJHU WUDQVLWLRQV >@ 3DWKZD\ f VKRZV D H9 KRW KROH LV JHQHUDWHG E\ $XJHU UHFRPELQDWLRQ RI D KRW HOHFWURQ ZLWK D WKHUPDO KROH 7KH QHLJKERULQJ KROH LV H[FLWHG WR KLJKHU HQHUJ\ DIWHU DEVRUELQJ WKH UHFRPELQDWLRQ HQHUJ\ )URP HQHUJ\ FRQVHUYDWLRQ WKH WKUHVKROG HOHFWURQ NLQHWLF HQHUJ\ QHHGHG WR JHQHUDWH D H9 KRW KROH LV AHOHFWURQ fWn;K a (* f f H9 H9 ZKHUH >f[BK LV WKH 6L6L KROH EDUULHU KHLJKW DQG (* LV WKH VLOLFRQ HQHUJ\ JDS 3DWKZD\ f VKRZV DQRWKHU KRW KROH JHQHUDWLRQ SURFHVV E\ $XJHU UHFRPELQDWLRQ RI D KRW HOHFWURQ ZLWK D WKHUPDO KROH ZLWK D GHHS YDOHQFH HOHFWURQ H[FLWHG LQWR WKH FRQGXFWLRQ EDQG HGJH WR FDUU\ DZD\ WKH UHFRPELQDWLRQ HQHUJ\ 7KH WKUHVKROG NLQHWLF HQHUJ\ RI WKH LQLWLDWLQJ .(H_HFWURQ UHTXLUHG WR JHQHUDWH D H9 KRW KROH LV JLYHQ E\

PAGE 128

f f AAHOHHWUDQk9 U .(H_HFWURQf§H9 s A AHOHFWURQ ‘ H 9 7f L (HZSH9 2IL p 4 .(KROH H9 2f p p f f p n f )LJXUH )RXU IXQGDPHQWDO H9 KRW KROH JHQHUDWLRQ SDWKZD\V LQLWLDWHG E\ D SULPDU\ HOHFWURQ .(HOHFWURQ LV WKH WKUHVKROG NLQHWLF HQHUJ\ RI WKH SULPDU\ HOHFWURQ f +RW KROH JHQHUDWLRQ E\ $XJHU UHFRPELQDWLRQ RI D KRW HOHFWURQ ZLWK D WKHUPDO KROH f +RW KROH JHQHUDWLRQ DQG WKHUPDO HOHFWURQ JHQHUDWLRQ E\ D $XJHU UHFRPELQDWLRQ RI D KRW HOHFWURQ ZLWK D WKHUPDO KROH f +RW KROH JHQHUDWLRQ E\ LQWHUEDQGLPSDFW RI D KRW HOHFWURQ ZLWK D WKHUPDO KROH WUDQVIHUULQJ WKH HQHUJ\ WR WKH WKHUPDO KROH f +RW KROH JHQHUDWLRQ DQG WKHUPDO HOHFWURQ JHQHUDWLRQ E\ LQWHUEDQGLPSDFW RI D KRW HOHFWURQ ZLWK D GHHS YDOHQFH EDQG HOHFWURQ

PAGE 129

.AHOHFWURQA f A;K (* B (* f H9 3DWKZD\ f VKRZV WKH KRW KROH JHQHUDWLRQ E\ LQWHUEDQG LPSDFW RI D KRW HOHFWURQ ZLWK D WKHUPDO KROH ZKHUH WKH KRW HOHFWURQ WUDQVIHUV DOO LWV NLQHWLF HQHUJ\ WR WKH WKHUPDO KROH )RU WKLV SDWKZD\ WKH WKUHVKROG NLQHWLF HQHUJ\ .(HOHFWURQ LV VLPSO\ WKH KROH EDUULHU KHLJKW ZKLFK LV JLYHQ E\ .(HOHFWURQ6 A;K f H9 3DWKZD\ f VKRZV WKH KRW KROH JHQHUDWLRQ E\ D KRW HOHFWURQ LPSDFWLQJ D GHHS YDOHQFH EDQG HOHFWURQ ZLWK D WKHUPDO HOHFWURQ WR UDLVH WKLV YDOHQFH HOHFWURQ WR WKH FRQGXFWLRQ EDQG HGJH )RU WKLV LQWHUEDQGLPSDFW SDWKZD\ WKH WKUHVKROG NLQHWLF HQHUJ\ .(HOHFWURQA LV JLYHQ E\ .(HOHFWURQA fWf;K (* f f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

PAGE 130

HQHUJLHV GHULYHG DQDO\WLFDOO\ EDVHG RQ HQHUJ\ DQG PRPHQWXP FRQVHUYDWLRQ DUH QHDUO\ HTXDO WR WKHLU WKHRUHWLFDO YDOXHV FRPSXWHG IURP HQHUJ\ FRQVHUYDWLRQ DORQH ZKLFK DUH GXH WR WKH HOHFWURQLQWHUYDOOH\ DQG XPNODSS WUDQVLWLRQV LQ VLOLFRQ >@ ([SHULPHQWV 7ZR JURXSV RI QFKDQQHO 026 WUDQVLVWRU VWUXFWXUHV ZLWK WZR GLIIHUHQW GUDZQ FKDQQHO OHQJWKV / _LP DQG SPf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f LV WKH PD[LPXP NLQHWLF HQHUJ\ JDLQHG E\ WKH FKDQQHO HOHFWURQ DFFHOHUDWLQJ IURP WKH ORZILHOG HGJH RI WKH GUDLQ MXQFWLRQ VSDFHFKDUJHOD\HU WR WKH HGJH RI WKH GUDLQ QS PHWDOOXUJLFDO MXQFWLRQ $IWHU HDFK VWUHVV LQWHUYDO WKH WUDQVLVWRU FKDUDFWHULVWLFV ZHUH PHDVXUHG XVLQJ WKH '&,9 PHWKRG LQ WKH GUDLQHPLWWHU FRQILJXUDWLRQ '('&,9f WR PRQLWRU WKH VWUHVVJHQHUDWHG 1,7 DQG 4RW

PAGE 131

5HVXOWV DQG 'LVFXVVLRQV 'UDLQ(PLWWHU &KDUDFWHULVWLF $ W\SLFDO IDPLO\ RI '('&,9 FXUYHV DW 9'6BVWUHVV 9 DQG 9*6VWUHVV 9 DQG VWUHVV WLPHV RI NVHF IURP D 6$0 UXQ WDNHQ RQ D _LP SP Q0267 LV VKRZQ LQ )LJXUH 7KH QHDUO\ VWDWLRQDU\ SHDN DW 9* 9 ZDV LGHQWLILHG E\ LWV H[ST9%'N%7f GHSHQGHQFH DV UHFRPELQDWLRQ RI QGUDLQMXQFWLRQ LQMHFWHG HOHFWURQV ZLWK WKH SERG\ KROHV DW WKH IDEULFDWLRQUHVLGXDO LQWHUIDFH ORFDWHG LQ WKH TXDVLQHXWUDO FKDQQHO UHJLRQ RXW RI WKH GUDLQMXQFWLRQ VSDFHFKDUJH UHJLRQ >@ ,WV JURZWK UDWH LV YHU\ ORZ EHFDXVH WKHUH DUH IHZ KRW HOHFWURQV DQG KROHV GXULQJ VWUHVV LQ WKH LQYHUWHG FKDQQHO RXWVLGH RI WKH GUDLQMXQFWLRQ VSDFHFKDUJHUHJLRQ 7KH IDVW JURZLQJ VHFRQGDU\ SHDN 9* 9 DIWHU NVHF VWUHVV ZDV LGHQWLILHG E\ LWV H[ST9'%N%7f GHSHQGHQFH VHH LQVHW RI )LJXUH f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

PAGE 132

)LJXUH 'UDLQHPLWWHU '&,9 '('&,9f FKDUDFWHULVWLFV RI /: LP ILP Q0267V PHDVXUHG DW AHEG 9 DQG 9%6 9 DIWHU &+( VWUHVVHG DW 9'6 9 9*6 9 DQG 9%6 9 % %RG\ ;f IRU N N N DQG N VHFRQGV 7KH LQVHW VKRZV WKH H[ST9%'QN%7f GHSHQGHQFH RI WKH VWUHVVLQGXFHG ERG\ FXUUHQW $,% ,E9JE 9 WVWUHVV NVHFf ,E9*E 9 WVWUHVV f LQGLFDWLQJ Q GXH WR UHFRPELQDWLRQ LQ WKH GUDLQ MXQFWLRQ VSDFHFKDUJH UHJLRQ

PAGE 133

<6& >eVL931T1$$f f >9SQ 9f1DDf@ :LWK 1DD [ %RURQFP WKHQ GXULQJ '('&,9 PHDVXUHPHQWV 9S1 9%,9%' f f 9 9 6XEVWLWXWLQJ 931 9 LQWR f WKH GUDLQ VSDFHFKDUJHOD\HU WKLFNQHVV GXULQJ '( '&,9 PHDVXUHPHQW LV HVWLPDWHG WR EH <6&PHV >HVL9T1$$fnr f >O9f[Of@QP QP 'XULQJ &+( VWUHVV DW 9'6 9 DQG 9*6 9 931 9'9'VDW f 9 9 9 WKHQ WKH GUDLQ VSDFHFKDUJHOD\HU WKLFNQHVV LV HVWLPDWHG WR EH HVL9T1$$fn f >O 9f[Of@QP QP %DVHG RQ WKHVH HVWLPDWLRQV WKH IDVWJURZLQJ LQWHUIDFH WUDSV DUH ORFDWHG LQ WKH FKDQQHO OHVV WKDQ QP IURP WKH QSVXUIDFH ERXQGDU\

PAGE 134

([SHULPHQWDO 27 %XLOGXS 7KUHVKROG 7KH SRVLWLYH R[LGH FKDUJH EXLOGXS FDQ EH PRQLWRUHG E\ WKH JDWHYROWDJH VKLIW RI WKH JURZLQJ ,% SHDN 9*%SHDN VKRZQ LQ )LJXUH VLQFH $427 r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fr,,F+VWUHVVA'UDLQfrGW f OTf},*+VWUHVVrWVWUHVVfA'UDLQ ZKHUH T LV WKH HOHFWURQLF FKDUJH [Of &RXORPEVf $'UDLQ LV WKH GUDLQ FRQWDFW DUHD DQG WVWUHVV LV WKH LQFUHPHQWDO VWUHVV WLPH $OO ILYH 9*%BSHDN FXUYHV DUH REVHUYHG WR VDWXUDWH DIWHU 1LQM ,2[,2 FP 1LQJ UHSRUWHG D ODUJH FDSWXUH FURVV VHFWLRQ DUHD aaFPff IRU KROH R[LGH WUDSV LQ 6L DQG REVHUYHG DQ DOPRVW b KROH WUDSSLQJ HIILFLHQF\ >@ 7KLV FRXOG SRVVLEO\ DFFRXQW IRU WKH 9*%fV VDWXUDWLRQ VKRZQ LQ )LJXUH ,Q )LJXUH WKH GDWD FRUUHVSRQGLQJ WR HDFK 9'6 VWUHVV YROWDJH ZDV OHDVWVTXDUH ILWWHG WR WZR H[SRQHQWLDO IXQFWLRQV WR H[WUDFW WKH 9*% SHDN YDOXH DW D FRQVWDQW KRW HOHFWURQ IOXHQFH RI 1LQM [ FPn 7KH ILWWHG UHVXOWV DUH VKRZQ DV VROLG OLQHV LQ

PAGE 135

)LJXUH .LQHWLFV RI SRVLWLYH R[LGH FKDUJH EXLOGXS YHUVXV VWUHVV IOXHQFH 1LQM LQ SP [ -LP Q0267V VWUHVVHG DW 9'6 9 9 9 DQG 9 DQG WKH VDPH 9*6 9

PAGE 136

)LJXUH )LJXUH VKRZV WKH H[WUDFWHG 9*%SHDN YDOXHV SORWWHG DV D IXQFWLRQ RI 9' 9'VDWf ZKLFK LV WKH DFFHOHUDWLRQ SRWHQWLDO WKURXJK WKH GUDLQMXQFWLRQ VSDFHFKDUJH OD\HU 9'VDW LV WKH SRWHQWLDO DFURVV WKH LQYHUWHG HOHFWURQ FKDQQHO DQG LV H[SHULPHQWDOO\ GHWHUPLQHG IURP WKH GUDLQ FRQGXFWDQFH PHDVXUHPHQW )RU WKH O22SP[ SP WUDQVLVWRU 9'VDW 9 ZDV REWDLQHG E\ OLQHDU H[WUDSRODWLQJ WKH J' GDWD WR J' $9 VKRZQ LQ )LJXUH 7KH GHDUWK RI GDWD VKRZQ LQ )LJXUH FRXOG QRW FOHDUO\ GHOLQHDWH WKH SUHVHQFH RI D VWUXFWXUH QHDU 9' 9'VDW 9 ZKLFK ZRXOG FRUUHVSRQG WR WKH HOHFWURQLPSDFW WKUHVKROG WUDQVLWLRQ f LQ )LJXUH f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f 7KH UHVXOWV VKRZQ LQ )LJXUHV DQG ZHUH FRPELQHG LQWR )LJXUH 5HDVRQDEOH DJUHHPHQW EHWZHHQ WKH / SP DQG / SP UHVXOWV LV REVHUYHG 7KH LQWHUVHFWLRQ RI WKH WZR GDVKHG OLQHV REWDLQHG YLD WZR VHSDUDWH OLQHDU/6) RI GDWD LQ WKH KLJK DQG ORZ YROWDJH UDQJH JLYHV 9' a9'VDW 9 ZKLFK DJUHHV ZLWK WKH WKHRUHWLFDO

PAGE 137

9' 9'VDWf 9f )LJXUH 3ORW RI SRVLWLYH FKDUJH EXLOGXS DV D IXQFWLRQ RI WKH HOHFWURQ NLQHWLF HQHUJ\ 9' 9'VDWf DW D VWUHVV IOXHQFH RI 1LQM [ FP IURP WKH GDWD JLYHQ LQ )LJXUH

PAGE 138

9 fD 9'6 9f )LJXUH ,'9' FKDUDFWHULVWLFV RI WKH LP[ 22ILP WUDQVLVWRU PHDVXUHG DW 9*6 9 7KH GUDLQ FRQGXFWDQFH J' LV GHWHUPLQHG IURP J' $,'$9' 7KH ]HUR 9'6LQWHUFHSW E\ WKH OLQHDU/6) OLQH RI J' JLYHV 9'VDW 9

PAGE 139

)LJXUH .LQHWLFV RI SRVLWLYH R[LGH FKDUJH EXLOGXS YHUVXV VWUHVV IOXHQFH 1LQM LQ SP [ SP Q0267V VWUHVVHG DW 9'6 9 9 9 9 9 9 9 9 DQG 9 DQG WKH VDPH 9*6 9

PAGE 140

9' 9'VDWf 9f )LJXUH 3ORW RI SRVLWLYH FKDUJH EXLOXS DV D IXQFWLRQ RI WKH HOHFWURQ NLQHWLF HQHUJ\ 9' 9'VDWf DW D VWUHVV IOXHQFH RI 1LQM [ FP IURP WKH GDWD JLYHQ LQ )LJXUH

PAGE 141

9' 9'VDWf 9f )LJXUH &RPSDULVRQ RI SRVLWLYH FKDUJH EXLOXS DV D IXQFWLRQ RI WKH HOHFWURQ NLQHWLF HQHUJ\ 9' 9'VDWf DW D VWUHVV IOXHQFH RI 1LQM [ FP IURP WKH GDWD JLYHQ LQ )LJXUHV DQG

PAGE 142

H9 LQWHUEDQGLPSDFW JHQHUDWLRQ WKUHVKROG GLVFXVVHG LQ 6HFWLRQ 7KXV KRW KROH LQMHFWLRQ LQWR 6L YLD WKH LQWHUEDQGLPSDFW SDWKZD\ GRPLQDWHV ZKHQ WKH HOHFWURQ NLQHWLF HQHUJ\ H[FHHGV WKH H9 WKUHVKROG LH T9' 9'VDWf H9 $VVXPLQJ 9'VDW 9*6 9*7 ZKHUH 9*7 LV WKH WKUHVKROG YROWDJH RI WKH Q0267 WKHQ D KLJK9' DQG ORZ9* &+(VWUHVV FRQGLWLRQ ZRXOG VDWLVI\ WKH 9' 9'VDWf 9' 9* 9*7f 9 WR SRVLWLYHO\ FKDUJH WKH 47 QHDU WKH GUDLQ VSDFHFKDUJHUHJLRQ E\ WKH &+(LPSDFWJHQHUDWHG KRW KROHV 7KLV FRXOG DFFRXQW IRU WKH KLJK9' DQG ORZ 9* VWUHVV FRQGLWLRQV XVHG LQ > @ WR LQMHFW KROHV LQWR WKH 6L GXULQJ &+( VWUHVV 27 %XLOGXS LQ 6KRUW&KDQQHO Q0267V )URP WKH ORQJFKDQQHO 0267Vf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

PAGE 143

6$0 H[SHULPHQW RQ WZR SP Q0267 VKRZQ LQ )LJXUH VWUHVVHG DW WZR WUDGLWLRQDO SRZHU VXSSO\ YROWDJHV 9'6VWUHVV 9 'HYLFH$f DQG 9 'HYLFH%f 7KLV VKRUW / SP 0267 LV IDEULFDWHG IURP D SP &026 WHFKQRORJ\ DQG KDV D QP WKLFN JDWH R[LGH DQG D GUDZQ FKDQQHO ZLGWK RI SP )RU WKLV / SP Q0267 9'VDW 9 LV GHWHUPLQHG IURP WKH FRQGXFWDQFH PHDVXUHPHQW )LJXUH SORWV WKH VWUHVVHGLQGXFHG $,%B'( OHIW \D[LVf YHUVXV 9*% RI 'HYLFH$ DQG 'HYLFH% VWUHVVHG DIWHU PLQ DQG PLQ UHVSHFWLYHO\ DQG WKH XQVWUHVVHG '( ,%9*% ULJKW \D[LVf FKDUDFWHULVWLFV RI 'HYLFH$ DQG 'HYLFH% GRWWHG OLQHf 7KH VWUHVVWLPH GLIIHUHQFH ZDV VHOHFWHG WR PDNH WKH VWUHVV GUDLQ IOXHQFH DOPRVW LGHQWLFDO LQ WKH WZR 6$0 WHVWV :LWK UHIHUHQFH WR WKH XQVWUHVVHG ,% SHDN DW 9*% 9 WKH VWUHVVLQGXFHG $,%B'( SHDNV IRU 'HYLFH$ DQG 'HYLFH% ERWK VKRZHG D QHJDWLYH VKLIW DORQJ WKH JDWH YROWDJH D[LV +RZHYHU 'HYLFH$ VWUHVVHG DW 9'6BVWUHVV 9 ZDV IRXQG WR KDYH ODUJHU EXLOGXS RI SRVLWLYH R[LGH FKDUJH 47 $9*% f§ 9 $4RW&R[f DQG LQWHUIDFH WUDSV 1,7 m $,%BSHDN a S$f &RPSDUDWLYHO\ 'HYLFH% VWUHVVHG DW 9'VVWUHVV 9 H[KLELWV D UDWKHU VPDOO 4R[ EXLOGXS DQG 1,7 JHQHUDWLRQ $9*% 9 DQG $,%BSHDN a S$f 7KLV FDQ EH H[SODLQHG E\ FRQVLGHULQJ WKH PDJQLWXGH RI 9' 9'VDWf ZKLFK LV WKH DFFHOHUDWLRQ YROWDJH IRU WKH FKDQQHO HOHFWURQ WKURXJK WKH GUDLQFKDQQHO VSDFHFKDUJHOD\HU DW WKH 6L6L LQWHUIDFH )RU 'HYLFH$ 9' 9'VDW 9 9 9 )RU 'HYLFH% 9' 9'VDW 9 9 9 7KH ODWWHU FRUUHVSRQGV WR WKH $XJHU WKUHVKROG RI 9 7KHVH WZR 6$0 UHVXOWV FRPSOHPHQW WKH SUHFHGLQJ ORQJ FKDQQHO UHVXOWV WKDW ZKHQ 9' 9'VDW ;K+ (*fT 9 WKH 4W EXLOGXS IURP EDUULHUVXUPRXQWLQJ KRW KROH LQMHFWLRQ GXULQJ &+(

PAGE 144

$,EGH $f 9*%9f )LJXUH 'UDLQ(PLWWHU '&,9 RI WZR _OP Q0267V 7UDQVLVWRU$ VWUHVVHG DW 9'6 9 IRU WVWUHVV PLQ DQG 7UDQVLVWRU% VWUHVVHG DW 9'6 9 IRU WVWUHVV PLQ 9*6VWUHVV 9 7KH GRWWHG OLQHV VKRZV WKH XQVWUHVVHG '('&,9 FKDUDFWHULVWLFV RI 7UDQVLVWRU$ DQG 7UDQVLVWRU% ,EGH $f

PAGE 145

VWUHVV ZLOO FHDVH ,W LV DOVR LQIHUUHG WKDW LQWHUIDFH WUDS ZRXOG EH D SULPDU\ GHJUDGDWLRQ IDFWRU DIIHFWLQJ WKH UHOLDELOLW\ RI ORZYROWDJH GHHS VXEPLFURQ 0267 ZKHUH 4W EXLOGXS LV QR ORQJHU LPSRUWDQW

PAGE 146

&+$37(5 5('8&7,21 2) ,17(5)$&( 75$36 ,1 3&+$11(/ 026 75$16,6725 '85,1* &+$11(/+27+2/( 675(66 ,QWURGXFWLRQ 7KH GRPLQDQW GHJUDGDWLRQ PHFKDQLVP LQ SFKDQQHO 026 WUDQVLVWRUV S0267Vf XQGHU FKDQQHO KRW KROH &++f VWUHVV DW ORZ 9*6 DQG KLJK 9'6 KDV EHHQ DWWULEXWHG WR EXLOGXS RI QHJDWLYH R[LGH FKDUJH 4TWf GXH WR HOHFWURQ WUDSSLQJ LQ WKH JDWH R[LGH QHDU WKH GUDLQ MXQFWLRQ VSDFHFKDUJH UHJLRQ > @ 5HOLDELOLW\ PRGHOLQJ RI S0267 KDV DVVXPHG FKDQQHOVKRUWHQLQJ FDXVHG E\ WKLV ORFDOL]HG QHJDWLYH R[LGH FKDUJH >@ )HZ VWXGLHV KDYH LQYHVWLJDWHG WKH PHFKDQLVPV RI LQWHUIDFH WUDS 1,7f JHQHUDWLRQ DQG LWV HIIHFWV RQ WKH SHUIRUPDQFH DQG GHJUDGDWLRQ RI S0267V > @ ,Q WKLV FKDSWHU ZH UHSRUW QRW RQO\ WKH JHQHUDWLRQ RI LQWHUIDFH WUDSV EXW DOVR UHGXFWLRQ RI LQWHUIDFH WUDSV IRU WKH ILUVW WLPH LQ S0267V GXULQJ &++ VWUHVV %DVHG RQ D YDULHW\ RI JHRPHWULFDO DQG ELDV GHSHQGHQFHV RQ WKH H[SHULPHQWDO UHVXOWV D K\GURJHQ UHDFWLRQWUDQVSRUW PRGHO LV FRQVWUXFWHG ([SHULPHQWDO 5HVXOWV 7KH WHVW WUDQVLVWRUV XVHG LQ WKH VWUHVVDQGPHDVXUH 6$0f H[SHULPHQWV ZHUH S0267V LQ QZHOO RQ SVXEVWUDWH ZLWK $ JDWH R[LGH DQG D GUDZQ DVSHFW UDWLR RI :/ SP -LP 7KH\ ZHUH VWUHVVHG DW URRP WHPSHUDWXUH XQGHU WKH FRQYHQWLRQDO &++ FRQGLWLRQ 9'6 9 9*6 9 9V[ 9 *HQHUDWLRQ DQG UHGXFWLRQ RI LQWHUIDFH WUDSV DQG FKDUJLQJ RI WKH R[LGH WUDSV DUH PRQLWRUHG E\ WKH '&,9 PHWKRG 7ZR

PAGE 147

ELDV FRQILJXUDWLRQV ZHUH HPSOR\HG LQ WKH '&,9 PHDVXUHPHQWV 'UDLQ(PLWWHU DQG 6RXUFH(PLWWHU IRUZDUG ELDV '('&,9 DQG 6('&,9f WR ORFDWH WKH LQWHUIDFH DQG R[LGH WUDSV 7KH WUDS ORFDWLRQV DUH IXUWKHU GHOLQHDWHG E\ WKH LGHDOLW\ IDFWRU RI WKH ,9 FKDUDFWHULVWLFV RI WKH IRUZDUGELDVHG MXQFWLRQ >@ )LJXUH VKRZV W\SLFDO '('&,9 FKDUDFWHULVWLFV RI D &++VWUHVVHG S0267 ZKHUH WKH VWUHVVJHQHUDWHG LQFUHPHQWDO EDVH FXUUHQW LV FRPSXWHG IURP H[SHULPHQWDO GDWD XVLQJ AVWUHVV A(%Of a A%:HVV } A(%Of a A%AVWUHVV f A(%Of f 7KH LQFUHPHQWDO UDWKHU WKDQ WRWDO ERG\ RU EDVHf FXUUHQW JLYHV YLVLELOLW\ RI WKH VPDOO &++VWUHVVLQGXFHG FKDQJHV VXFK DV WKH GHFUHDVLQJ SHDN DW 9*% 9 LQ )LJXUHV Ef DQG Gf DQG WKH LQFUHDVLQJWKHQGHFUHDVLQJ SHDN DW 9*% 9 LQ )LJXUH Ff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f IRU WKH WZR ELDV FRQILJXUDWLRQV '( DQG 6(f VKRZQ LQ )LJXUHV DfGf 7KHVH IRXU '&,9 SORWV KHOS WR ORFDWH WKH LQWHUIDFH DQG R[LGH WUDSV $W WKH ORZHU IRUZDUG ELDV 9f $,% LV GRPLQDWHG E\ HOHFWURQKROH UHFRPELQDWLRQ DW WKH LQWHUIDFH WUDSV ORFDWHG RYHU WKH VXUIDFH VSDFHFKDUJHUHJLRQ RI WKH IRUZDUGELDVHG GUDLQ RU VRXUFH

PAGE 148

,9 928f D,9 L L L L L L L L L L L L L L L , 6(9VE 9 R VHF r VHF r VHF r R VHF } VHF ; VHF L L L L L L L L L L L L 't9'E 9 QB N n 7 , _ L _ 6(9VE 9 VHF } VHF r VHF f§ f§ } VHF B R VHF R VHF } VHF D VHF R VHF Q r VHF Qf§ LLULLQPQUQLLLLQPQLLfLLSULLIL_@nQLLI Eff LQ O N Gf ‘ ‘ n ‘ ‘ ‘ L ‘ n ‘ L L ‘ ‘ L L } ‘ ‘ } 9*%9f 9TE9f )LJXUH 7KH P9 DQG P9 'UDLQ (PLWWHU '(f DQG 6RXUFH (PLWWHU 6(f '&,9 FKDUDFWHULVWLFV RI VLOLFRQ S0267V ZLWK :/ _LPSP VKRZLQJ WKH 9*% GHSHQGHQFH RI $,% DW 9&% DQG WZR PHDVXUHPHQW IRUZDUG ELDVHV 9HE 9 DQG 9f DIWHU FKDQQHO KRW KROH VWUHVVHG DW 9'6 9 9*6 9 DQG 9V[ 9 IRU DQG VHFRQGV

PAGE 149

MXQFWLRQ >@ $W WKH KLJKHU IRUZDUG ELDV 9f $,% LV GRPLQDWHG E\ HOHFWURQKROH UHFRPELQDWLRQ DW WKH LQWHUIDFH WUDSV ORFDWHG RYHU WKH VXUIDFH FKDQQHO UHJLRQ EHWZHHQ WKH VRXUFH DQG WKH GUDLQ >@ 7KH ,9 FKDUDFWHULVWLFV RI WKH IRUZDUGELDVHG GUDLQHPLWWHU '(f DQG VRXUFH HPLWWHU 6(f MXQFWLRQV IXUWKHU KHOS WR ORFDWH WKH LQWHUIDFH WUDSV 7KH UHFLSURFDO VORSH RU LGHDOLW\ IDFWRU Q LQ $,% m H[ST9(%QN7f LV UHDG IURP )LJXUH DQG FRSLHG WR )LJXUH 7KH YDOXH Q VLJQLILHV UHFRPELQDWLRQ LQ WKH VSDFHFKDUJH UHJLRQ ZKLOH Q O LQ WKH TXDVLQHXWUDO FKDQQHOf UHJLRQ >@ ,W LV HYLGHQW IURP WKH '('&,9 RI )LJXUHV Df DQG Ef WKDW WKH UDSLGO\ JURZLQJ SHDN DW 9*% 9 ZLWK Q RULJLQDWHV IURP &++JHQHUDWHG LQWHUIDFH WUDSV ORFDWHG RYHU WKH SGUDLQQZHOO VXUIDFH VSDFHFKDUJHUHJLRQ 7KH KLJK JURZWK UDWH UHIOHFWV WKH KLJK GHQVLW\ RI SULPDU\ KRW KROHV DQG VHFRQGDU\ KRW HOHFWURQV LQ WKLV UHYHUVHELDVHG VSDFHFKDUJHUHJLRQ 7KH ORFDWLRQ RI WKH &++JHQHUDWHG LQWHUIDFH WUDSV LV IXUWKHU FRQILUPHG E\ WKH 6('&,9 RI )LJXUH Ff DW 96% 9 ZKLFK VKRZV D UHVLGXDOIDEULFDWLRQ LQWHUIDFH WUDSV RYHU WKH SVRXUFHQZHOO VXUIDFH VSDFHFKDUJHUHJLRQ EHFDXVH Q f WKDW LV RYHUZKHOPHG E\ WKH PLGFKDQQHO UHFRPELQDWLRQ Q Of FRPSRQHQW ZKHQ 96% LV LQFUHDVHG WR 9 LQ )LJXUH Gf 7KH VKLIW RI WKH Q SHDN LQ WKH SRVLWLYH 9*% GLUHFWLRQ DV LQGLFDWHG LQ )LJXUHV Df DQG Ef VKRZV WKH QHJDWLYH FKDUJLQJ RI WKH R[LGH WUDSV RYHU WKH GUDLQMXQFWLRQ VSDFHFKDUJH UHJLRQ E\ EDUULHU VXUPRXQWLQJ 6L6L -![H H9f VHFRQGDU\ KRW HOHFWURQV JHQHUDWHG E\ WKH SULPDU\ RU FKDQQHO KRW KROHV LQ WKH SGUDLQ VLPLODU WR WKDW H[SHULPHQWDOO\ SURYHQ IRU SRVLWLYH FKDUJLQJ RI WKH R[LGH WUDSV GXULQJ &+( VWUHVV LQ Q0267 SUHVHQWHG LQ &KDSWHU

PAGE 150

$,E$f $,E$f '( WVWUHVV N6 L U $,E r 9JE 9 r 9*% 9 f 8 9JU 9 f 9(%9f VH AVWUHVV N6 A W B ’ 9JE 9 % R 9JE 9 0 f§ r 9JE 9 f§ 9 f§ J U -;n Ur 0+ A Z M E7 )LJXUH 7KH IRUZDUG FXUUHQWYROWDJH FKDUDFWHULVWLFV $,% YV 9(%f RI WKH GUDLQ DQG VRXUFH MXQFWLRQ RI WKH DVSHFWUDWLR S0267V LQ WKH Df '( DQG Ef 6( PHDVXUHPHQW FRQILJXUDWLRQV DW WKUHH JDWH ELDVHV 9*% 9 9 DQG 9f DIWHU WKH &++VWUHVVHG IRU NV 7KH WZR EURNHQ OLQHV JLYH WKH WKHRUHWLFDO UHFLSURFDO VORSHV Q O DQG Q

PAGE 151

7KH QHJDWLYH SHDN VWDUWLQJ DW 9*% 9 LQ )LJXUHV Ef DQG Gf RULJLQDWHV IURP WKH GHQVLW\UHGXFWLRQ RI WKH UHVLGXDO IDEULFDWLRQ LQWHUIDFH WUDSV ORFDWHG LQ WKH FKDQQHO GXULQJ &++ VWUHVV ,WV PLGFKDQQHO ORFDWLRQ LV GHGXFHG IURP LWV Q O LGHDOLW\ IDFWRU SUHVHQW DW WKH KLJKHU IRUZDUG ELDV 96% 9f WKDW FDXVHV UHFRPELQDWLRQ LQ WKH TXDVLQHXWUDO UHJLRQ VLJQLILHG E\ Q Of WR GRPLQDWH DQG WKH QHDU V\PPHWU\ EHWZHHQ 6( DQG '( 7KH VOLJKW DV\PPHWU\ RU WKH KLJKHU UHGXFWLRQ UDWH LQ WKH '( FRQILJXUDWLRQ )LJXUH Ef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m+ DQG 6L2m+ DW WKH 6L6L LQWHUIDFH RYHU WKH GUDLQMXQFWLRQ VSDFHFKDUJH UHJLRQ LV UHOHDVHG E\ WKH HQHUJHWLF VHFRQGDU\ HOHFWURQV JHQHUDWHG E\ WKH SULPDU\ FKDQQHO KRW KROHV )LJXUHV Df DQG Ef VKRZ WKH NLQHWLFV RI WKH JURZWK RI WKH LQWHUIDFH WUDSV RYHU LQ WKH GUDLQMXQFWLRQ VSDFHFKDUJH UHJLRQ Q f ZKLFK LV LQLWLDOO\ OLPLWHG E\ K\GURJHQERQG EUHDNLQJ ZLWK WKH OLQHDU VWUHVV WLPH GHSHQGHQFH DQG WKHQ OLPLWHG E\ K\GURJHQ GLIIXVLRQ ZLWK VTXDUHURRW VWUHVV WLPH GHSHQGHQFH ,W VDWXUDWHV ZKHQ DOO WKH K\GURJHQDWHG VLOLFRQ DQG R[\JHQ GDQJOLQJ ERQGV DUH GHK\GURJHQDWHG

PAGE 152

&' Q W L D Dn L UQ L U 0+ VDWXUDWHV $$OLL $ $ f§ L2K W9 $ '&,9 9GE 9 Y&% RY $ '('&,9 9JE 9 Df LnnO //O/ , /O, /BX, / )LJXUH ,QWHUIDFH WUDS JHQHUDWLRQ NLQHWLFV REWDLQHG E\ UHJXODU DQG IDVW '&,9 VDPSOLQJ RI VWUHVVHG S0267V E\ FKDQQHO KRW KROHV DW 9'6 9 9*6 9 DQG 9V[ 9 Df )DVW '&,9 VDPSOLQJ DW 9'% 9 DQG 9*% 9 Ef 5HJXODU '&,9 DW 9(% 9 DW 9*% DQG 9

PAGE 153

7KH GHFUHDVLQJ SHDN VWDUWLQJ DW 9*% 9 ZKLFK LV SORWWHG DW 9*% 9 LQ )LJXUH IROORZV D VTXDUHURRW VWUHVVWLPH GHSHQGHQFH VXJJHVWLQJ WKDW WKH K\GURJHQ GLIIXVLRQ LV WKH UDWHOLPLWLQJ VWHS LQ WKH K\GURJHQDWLRQ RI WKH LQWHUIDFH WUDSV LQ WKH PLGn FKDQQHO UHJLRQ 7KLV LV FRQVLVWHQW ZLWK WKH GLIIXVLRQ GHOD\ RI WKH K\GURJHQ UHOHDVHG E\ FKDQQHO KRW KROHV IURP K\GURJHQDWHG ERURQ 6L%}+}6L LQ WKH SGUDLQ %RURQ K\GURJHQDWLRQ ZDV ILUVW LGHQWLILHG DQG YHULILHG E\ 6DK6XQ7]RXR LQ D VHULHV RI 026 FDSDFLWRU DYDODQFKH H[SHULPHQWV >@ 7KUHH DGGLWLRQDO REVHUYDWLRQV VXSSRUWHG WKLV PRGHO f WKH JURZWK UDWH RI WKH GHFUHDVLQJ SHDN DW 9*% 9 LV PXFK ORZHU LQ ORQJFKDQQHO S0267V WKDQ LQ VKRUW FKDQQHOV f D KLJK GHQVLW\ RI K\GURJHQDWHG ERURQ LV DYDLODEOH LQ WKH KHDYLO\GRSHG SGUDLQ >D b ERURQ K\GURJHQDWLRQ LQ WKH KHDYLO\ GRSHG SGUDLQ a %RURQFPf FRQWDLQV a ERURQK\GURJHQ ERQGV@ DQG f WKH GHFUHDVLQJ SHDN LV EDUHO\ QRWLFHDEOH LQ Q0267V ZKLFK VXJJHVWHG WKDW WKH VRXUFH RI WKH K\GURJHQ ZDV IURP WKH K\GURJHQDWHG ERURQDFFHSWRU LQ Q0267fV SZHOO ZKLFK KDV D PXFK ORZHU ERURQ FRQFHQWUDWLRQ WKDQ WKH SGUDLQ RI WKH S0267 )LJXUH LOOXVWUDWHV WKH K\GURJHQ GLIIXVLRQ SDWKZD\V IRU WKH &++VWUHVVHG S0267 +\GURJHQ UHOHDVHG IURP WKH EURNHQ 6Lm DQG 6L2 ERQGV DQG GHK\GURJHQDWHG ERURQ DFFHSWRUV 6L%} FDQ HLWKHU GLIIXVH WRZDUGV WKH SSRO\6L DORQJ WKH 6LQ6L FKDQQHO RU LQ WKH Q6L VXEVWUDWH +RZHYHU K\GURJHQ GLIIXVLRQ WRZDUGV WKH SSRO\6L GRHV QRW SDUWLFLSDWH LQ WKH K\GURJHQDWLRQ SURFHVV DW WKH 6L6L LQWHUIDFH WKRXJK LW RFFXUV DW D IDVWHU UDWH WKDQ LWV ODWHUDO GLIIXVLRQ WRZDUGV WKH PLGFKDQQHO UHJLRQ DV VKRZQ E\ WKH IROORZLQJ HVWLPDWLRQ XVLQJ /'LII 9'Wf

PAGE 154

ULUUP Q PL L UPPL P Q Q S6XEVWUDWH 6XEVWUDWH )LJXUH &URVVVHFWLRQDO YLHZ RI D S0267 XQGHU &++ VWUHVV 7KH K\GURJHQ GLIIXVLRQ SDWKZD\V IURP WKH EURNHQ LQWHUIDFLDO ERQGV DQG GHK\GURJHQDWHG ERURQ DUH VKRZQ 7KH GRWWHG OLQH LQGLFDWHV WKH HGJH RI WKH VSDFHFKDUJH OD\HU

PAGE 155

f f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

PAGE 156

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f FKDUDFWHUL]DWLRQ WHFKQLTXH 7KLV WHFKQLTXH HPSOR\V WKH VXUIDFH UHFRPELQDWLRQ FXUUHQW RI WKH SDUDVLWLF JDWHFRQWUROOHG %-7 LQ 0267 VWUXFWXUH DV D GHJUDGDWLRQ PRQLWRU 7ZR DSSOLFDWLRQV DUH LOOXVWUDWHG f VHSDUDWLRQ RI VWUHVVJHQHUDWHG R[LGH DQG LQWHUIDFH WUDSV LQ VXEPLFURQ 0267V DQG f VSDWLDO SURILOLQJ RI WKH VWUHVVJHQHUDWHG LQWHUIDFH WUDS GLVWULEXWLRQ 7KH VWUHVVJHQHUDWHG VXUIDFH UHFRPELQDWLRQ FXUUHQW LV VKRZQ WR H[KLELW D OLQHDU FRUUHODWLRQ ZLWK WKH WUDGLWLRQDOO\ PHDVXUHG WUDQVLVWRUfV SDUDPHWHUV ,'VDW 9*7 DQG 6f $ VLPSOH GUDLQFXUUHQW GHJUDGDWLRQ PRGHO ZDV VXFFHVVIXOO\ XVHG WR DFFRXQW IRU WKH REVHUYHG FRUUHODWLRQ 7KLV FRUUHODWLRQ SURYLGHV D EDVLV IRU XVLQJ WKH '&,9 PHWKRG WR

PAGE 157

PRQLWRU 026 WUDQVLVWRU GHJUDGDWLRQ $ WXUQNH\ '&,9 VRIWZDUH ZDV GHYHORSHG DQG GHOLYHUHG WR PHPEHUV RI WKH 6HPLFRQGXFWRU 5HVHDUFK &RUSRUDWLRQ $ SK\VLFEDVHG 7LPHWR)DLOXUH H[WUDFWLRQ DOJRULWKP ZDV GHPRQVWUDWHG XVLQJ WKH FXUUHQWDFFHOHUDWLRQ &+& VWUHVV WHFKQLTXH 7KH FXUUHQW DFFHOHUDWLRQ PHWKRG XVHV D IRUZDUGELDVHG SQ MXQFWLRQ WR LQFUHDVH WKH FKDQQHO KRW FXUUHQW WR HQKDQFH WKH WUDQVLVWRUf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fV ERQGEUHDNLQJ LQWHUIDFH WUDS JHQHUDWLRQ WKHRU\ DQG VKRZHG WKDW LQWHUIDFH WUDSV JHQHUDWHG GXULQJ &+& VWUHVV KDV D H9 WKUHVKROG HQHUJ\ 6LOLFRQK\GURJHQ 6L+f ERQGV DW WKH 6L6L LQWHUIDFH DUH FRQFOXGHG WR EH EURNHQ E\ FKDQQHO KRW KROHV LQ Q0267 DQG E\ FKDQQHO KRW HOHFWURQV LQ S0267 GXULQJ KLJKYROWDJH VWUHVV $ 9 WKUHVKROG LQWHUIDFH WUDS JHQHUDWLRQ YLD $XJHU UHFRPELQDWLRQ LV LGHQWLILHG WR EH D GRPLQDQW GHJUDGDWLRQ PHFKDQLVP IRU ORZYROWDJH 9 DQG EHORZf 0267 'XULQJ &+( VWUHVV R[LGH KROH WUDSV RYHU WKH GUDLQ MXQFWLRQ VSDFHFKDUJH OD\HU DUH FKDUJHG E\ FDSWXULQJ KROHV LQMHFWHG LQWR WKH R[LGH RYHU WKH 6L6L KROH EDUULHU )RXU

PAGE 158

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

PAGE 159

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

PAGE 160

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

PAGE 161

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f ,QVXODWHG *DWH )LHOG(IIHFW 7UDQVLVWRU ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO(' 1R SS >@ --6DQFKH] .. +VXHK 7$ 'H0DVVD 'UDLQ(QJLQHHUHG +RW(OHFWURQ 5HVLVWDQW 'HYLFH 6WUXFWXUHV $ 5HYLHZ ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SSO -XQ >@ )& +VX +RW&DUULHU5HVLVWDQW 6WUXFWXUHV &KDSWHU 9/6, (OHFWURQLFV 0LFURVWUXFWXUH 6FLHQFH 9RO $GYDQFHG 026 'HYLFH 3K\VLFV $FDGHPLF 3UHVV

PAGE 162

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f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

PAGE 163

>@ 7+ 1LQJ DQG +1 @ : 6KRFNOH\ 3UREOHPV 5HODWHG WR SQ -XQFWLRQV LQ 6LOLFRQ 6ROLG6WDWH (OHFWURQLFV 9RO 1RO SS >@ 7+ 1LQJ &0 2VEXUQ DQG +1 @ & +X /XFN\(OHFWURQ 0RGH RI &KDQQHO +RW (OHFWURQ (PLVVLRQ ,('0 SS >@ 6 7DP 3. .R DQG & +X /XFN\(OHFWURQ 0RGHO RI &KDQQHO +RW(OHFWURQ ,QMHFWLRQ LQ 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO (' 1R SS >@ 6$ $EEDV DQG 5& 'RFNHUW\ +RW&DUULHU ,QVWDELOLW\ LQ ,*)(7fV $SSO 3K\V /HWW 9RO 1R SS $XJ >@ 3( &RWWUHO 5 7URXWPDQ DQG 7+ 1LQJ +RW(OHFWURQ (PLVVLRQ LQ 1 &KDQQHO ,*)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR SS $SU >@ .. 1J DQG *: 7D\ORU (IIHFWV RI +RW&DUULHU 7UDSSLQJ LQ Q DQG S &KDQQHO 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS $XJY >@ % 'R\OH 0 %RXUFHULH -& 0DUFKHWDX[ DQG $ %RXGRX ,QWHUIDFH 6WDWH &UHDWLRQ DQG &KDUJH 7UDSSLQJ LQ WKH 0HGLXPWR+LJK 9ROWDJH 5DQJH 9G 9J 9Gf 'XULQJ +RW&DUULHU 6WUHVVLQJ RI Q026 7UDQVLVWRUV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS 0DU >@ -< &KRL 3. 3R & +X DQG : 6FRWW+RW &DUULHU,QGXFHG 'HJUDGDWLRQ RI 0HWDO2[LGH6HPLFRQGXFWRU )LHOG (IIHFW 7UDQVLVWRUV 2[LGH &KDUJH 9HUVXV ,QWHUIDFH 7UDSV $SSO 3K\V 9RO 1RO SS -DQ >@ 6$ $EEDV 6XEVWUDWH &XUUHQW $ 'HYLFH DQG 3URFHVV 0RQLWRU ,('0 SS 'HF >@ 3DXO +HUHPDQV 5XGL %HOOHQV *XLGR *URHVHQHNHQ DQG +HUPDQ ( 0DHV &RQVLVWHQW 0RGHO IRU WKH +RW&DUULHU 'HJUDGDWLRQ LQ QFKDQQHO DQG S &KDQQHO 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS 'HF

PAGE 164

>@ 7< &KDQ 3. 3R DQG & +X 'HSHQGHQFH RI &KDQQHO (OHFWULF )LHOG RQ 'HYLFH 6FDOLQJ ,((( (OHFWURQ 'HY /HWW 9RO 1R SS )HE >@ < 1LVVDQ&RKHQ $ 1RYHO )ORDWLQJ*DWH 0HWKRG IRU 0HDVXUHPHQW RI 8OWUD /RZ +ROH DQG (OHFWURQ *DWH &XUUHQWV LQ 026 7UDQVLVWRUV ,((( (OHFWURQ 'HY /HWW 9RO 1R SS 0DU >@ 16 6DNV 3/ +HUHPDQV /9' +RYH +( 0DHV 5)' .HHUVPDHFNHU DQG *'HFOHUFN 2EVHUYDWLRQ RI +RW+ROH ,QMHFWLRQ LQ 1026 7UDQVLVWRUV 8VLQJ D 0RGLILHG )ORDWLQJ*DWH 7HFKQLTXH ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR SS 0D\ >@ &KLK7DQJ 6DK (IIHFW RI 6XUIDFH 5HFRPELQDWLRQ DQG &KDQQHO RQ 31 -XQFWLRQ DQG 7UDQVLVWRU &KDUDFWHULVWLFV ,5( 7UDQVDFWLRQ RQ (OHFWURQ 'HYLFHV 9RO 1RO SS -DQ >@ &KLK7DQJ 6DK $ 1HZ 6HPLFRQGXFWRU 7HWURGH 7KH 6XUIDFH3RWHQWLDO &RQWUROOHG 7UDQVLVWRU 3URFHHGLQJ RI WKH ,5( 9RO 1R SS 1RY >@ -7 .DYDOLHURV DQG &7 6DK 6HSDUDWLRQ RI ,QWHUIDFH DQG 1RQXQLIRUP 2[LGH 7UDSV E\ WKH '& &XUUHQW9ROWDJH 0HWKRG ,((( 7UDQVDFWLRQ RQ (OHFWURQ 'HYLFHV 9RO 1RO SS -DQ >@ $ 1HXJURVFKHO &7 6DK DQG 0 &DUUROO 'HJUDGDWLRQ RI %LSRODU 7UDQVLVWRU &XUUHQW *DLQ %\ +RW +ROHV 'XULQJ 5HYHUVH (PLWWHU%DVH %LDV 6WUHVV ,((( 7UDQVDFWLRQ RQ (OHFWURQ 'HYLFHV 9RO 1R SS $XJ >@ &7 6DK $ 1HXJURVFKHO .0 +DQ DQG '2 0DUWLQ 7KH '& &XUUHQW 9ROWDJH 0HWKRG IRU 0HDVXULQJ ,QWHUIDFH 7UDSV LQ 6XEPLFURQ 6LOLFRQ 7UDQVLVWRUV $ 'HPRQVWUDWLRQ 7(&+&21f 3KRHQL[ $UL]RQD 86$ >@ $UQRVW 1HXJURVFKHO DQG &KLK7DQJ 6DK )XQGDPHQWDO 3K\VLFDO 'HJUDGDWLRQ 0HFKDQLVPV LQ 6LOLFRQ %LSRODU -XQFWLRQ 7UDQVLVWRUV 3XEOLFDWLRQ & 6HPLFRQGXFWRU 5HVHDUFK &RUSRUDWLRQ 5HVHDUFK 7ULDQJOH 3DUN 1& >@ &7 6DK 51 1R\FH DQG : 6KRFNOH\ &DUULHU *HQHUDWLRQ DQG 5HFRPELQDWLRQ LQ 31 -XQFWLRQV DQG 31 -XQFWLRQ &KDUDFWHULVWLFV 3URFHHGLQJ RI WKH ,5( 9RO SS 6HS >@ 0 : +LOOHQ DQG +ROVEULQN 7KH %DVH &XUUHQW 5HFRPELQLQJ DW WKH R[LGL]HG VLOLFRQ VXUIDFH 6ROLG6WDWH (OHFWURQLFV YRO SS 0D\

PAGE 165

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f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f RQ S >@ '5 @ + +DGGDUD DQG 6 &ULVWRORYHDQX 7ZR'LPHQVLRQDO 0RGHOLQJ RI /RFDOO\ 'DPDJHG 6KRUW&KDQQHO 026)(7fV 2SHUDWLQJ LQ WKH /LQHDU 5HJLRQ ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR SS )HE

PAGE 166

>@ $QGUHDV 6FKZHULQ :LOIULHG +DQVFK DQG :HUQHU :HEHU 7KH 5HODWLRQVKLS %HWZHHQ 2[LGH &KDUJH DQG 'HYLFH 'HJUDGDWLRQ $ &RPSDUDWLYH 6WXG\ RI Q DQG S &KDQQHO 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR 'HF >@ 7 0L]XQR 7 .RYRUL < 6DLWRK 6 6DZDGD DQG 7 7DQDND *DWH)ULQJLQJ )LHOG (IIHFWV RQ +LJK 3HUIRUPDQFH LQ +LJK 'LHOHFWULF /'' 6SDFHU 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR SS $SU >@ 7 :DQJ & +XDQJ 3& &KRX 6WHYH 66 &KXQJ DQG 7( &KDQJ (IIHFWV RI +RW &DUULHU ,QGXFHG ,QWHUIDFH 6WDWH *HQHUDWLRQ LQ 6XEPLFURQ /'' 026)(7f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fV ,((( (OHFWURQ 'HYLFH /HWWHUV YROOO QR SS )HE >@ 0DULR $QFRQD 1HOVRQ 6 6DNV DQG 'DQLHO 0F&DUWK\ /DWHUDO 'LVWULEXWLRQ RI +RW&DUULHU,QGXFHG ,QWHUIDFH 7UDSV LQ 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR 'HF >@ 3DXO +HUHPDQV -RKQ :LWWHUV *XLGR *URHVHQHNHQ DQG +HUPDQ ( 0DHV $QDO\VLV RI WKH &KDUJH 3XPSLQJ 7HFKQLTXH DQG ,WV $SSOLFDWLRQV IRU WKH (YDOXDWLRQ RI 026)(7 'HJUDGDWLRQ ,((( 7UDQV (OHFWURQ 'HYLFHV YRO QR SS -XO

PAGE 167

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

PAGE 168

>@ 9*. 5HGGL ,QIOXHQFH RI 6XUIDFH &RQGLWLRQV RQ 6LOLFRQ 3ODQDU 7UDQVLVWRU &XUUHQW *DLQ 6ROLG6WDWH (OHFWURQLFV YRO SS >@ &KLK7DQJ 6DK )XQGDPHQWDO RI VROLGVWDWH HOHFWURQLFV :RUOG 6FLHQWLILF 3XEOLVKLQJ &R 6LQJDSRUH 6HH (TQf RQ S >@ &KLK7DQJ 6DK )XQGDPHQWDO RI VROLGVWDWH HOHFWURQLFV :RUOG 6FLHQWLILF 3XEOLVKLQJ &R 6LQJDSRUH 6HH f IRU ,'BVDW DQG f DQG f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fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS 0DU >@ ( 7DNHGD $ 6KLPL]X DQG 7 +DJLZDUD 5ROH RI +RW+ROH ,QMHFWLRQ LQ +RW &DUULHU (IIHFWV DQG WKH 6PDOO 'HJUDGHG &KDQQHO 5HJLRQ LQ 026)(7fV ,((( (OHFWURQ 'HYLFHV /HWWHUV YRO QR SS 6HS >@ + *HVFK -3 /HEXUWRQ DQG 'RUGD *HQHUDWLRQ RI ,QWHUIDFH 6WDWHV E\ +RW +ROH ,QMHFWLRQ LQ 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS 0D\ >@ 5 )DLU DQG 5& 6XQ 7KUHVKROG9ROWDJH ,QVWDELOLW\ LQ 026)(7fV 'XH WR &KDQQHO+RW +ROH (PLVVLRQ ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1RO SS -DQ >@ 6. /DL 7ZR&DUULHU 1DWXUH RI ,QWHUIDFH6WDWH *HQHUDWLRQ LQ +ROH 7UDSSLQJ DQG 5DGLDWLRQ 'DPDJH $SSO 3K\V /HWW 9RO 1RO SS -XO\ >@ 6. /DL ,QWHUIDFH 7UDS *HQHUDWLRQ LQ 6LOLFRQ 'LR[LGH :KHQ (OHFWURQV $UH &DSWXUHG E\ 7UDSSHG +ROHV $SSO 3K\V 9RO 1R SS 0D\

PAGE 169

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f >@ 3ULYDWH &RPPXQLFDWLRQ RI &7 6DK WR 0 +DQ RQ -XQH DQG $ 1HXJURVFKHO RQ -XQH >@ .DUO 5 +RIPDQ &KULVWRSK :HUQHU :HUQHU :HEHU DQG *HUKDUG 'RUGD +RW (OHFWURQ DQG +ROH (PLVVLRQ (IIHFWV LQ 6KRUW Q&KDQQHO 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS 0DU >@ 3 +HUHPDQV +( 0DHV DQG 1 6DNV (YDOXDWLRQ RI +RW &DUULHU 'HJUDGDWLRQ RI 1&KDQQHO 026)(7fV ZLWK WKH &KDUJH 3XPSLQJ 7HFKQLTXH ,((( (OHFWURQ 'HYLFH /HWWHUV 9RO ('/ 1R SS -XO

PAGE 170

>@ (LML 7DNHGD &DU\ < @ :HQOLDQJ &KHQ DQG 7VR3LQJ 0D 2[LGH &KDUJH %XLOGXS DQG 6SUHDGRXW 'XULQJ &KDQQHO+RW&DUULHU ,QMHFWLRQ LQ 1026)(7fV ,((( (OHFWURQ 'HYLFH /HWWHUV 9RO 1R SS -XQ >@ :HUQHU :HEHU 0DUWLQ %UR[ 5RODQG 7KHZV DQG 1HOVRQ 6DNV +RW+ROH ,QGXFHG 1HJDWLYH 2[LGH &KDUJH LQ Q026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS $XJ >@ 6HH )LJXUH % RQ S DQG 7DEOH % RQ S RI &KLK7DQJ 6DK )XQGDPHQWDO RI 6ROLG6WDWH (OHFWURQLF 6WXG\ *XLGH :RUOG 6FLHQWLILF 6LQJDSRUHf >@ 7+ 1LQJ &DSWXUH FURVV VHFWLRQ DQG WUDS FRQFHQWUDWLRQ RI KROHV LQ VLOLFRQ GLR[LGH -RXUQDO RI $SSOLHG 3K\VLF 9RO 1R SS 0DU >@ 6HH 6HFWLRQ SS RI &KLK7DQJ 6DK )XQGDPHQWDO RI 6ROLG6WDWH (OHFWURQLF :RUOG 6FLHQWLILF 6LQJDSRUHf >@ (LML 7DNHGD @ $QGUHDV 6FKZULQ DQG :LOIUHG +DQVFK DQG :HUQHU :HEHU 7KH 5HODWLRQVKLS %HWZHHQ 2[LGH &KDUJH DQG 'HYLFH 'HJUDGDWLRQ $ &RPSDUDWLYH 6WXG\ RI Q DQG S&KDQQHO 026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS 'HF >@ 7RQJ&KHUQ 2QJ 3LQJ.HXQJ .R DQG &KHQPLQJ +X +RW&DUULHU &XUUHQW 0RGHOLQJ DQG 'HYLFH 'HJUDGDWLRQ LQ 6XUIDFH&KDQQHO S026)(7fV ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS -XO >@ ) 0DWVXRND + ,ZDL + +D\DVKLGD +DPD 7 7R\RVKLPD DQG .HQML 0DHJXFKL $QDO\VLV RI +RW&DUULHU,QGXFHG 'HJUDGDWLRQ 0RGH RQ S026)(7V ,((( 7UDQV (OHFWURQ 'HYLFHV 9RO 1R SS -XQ >@ 7RVKLDNL 7VXFKL\D
PAGE 171

>@ &7 6DK -DFN <& 6XQ DQG -RVHSK -7 7]RX *HQHUDWLRQ$QQHDOLQJ .LQHWLFV DQG $WRPLF 0RGHOV RI D &RPSHQVDWLQJ 'RQRU LQ WKH 6XUIDFH 6SDFH &KDUJH /D\HU RI 2[LGL]HG 6LOLFRQ $SSO 3K\V YRO 1R SS )HE >@ &7 6DK -DFN <& 6XQ DQG -RVHSK -7 7]RX 'HDFWLYDWLRQ RI WKH %RURQ $FFHSWRU LQ 6LOLFRQ E\ +\GURJHQ $SSO 3K\V /HWW YRO QR SS -XO >@ &7 6DK 6DPXHO &6 3DQ DQG &KDUOHV &+ +VX +\GURJHQDWLRQ DQG $QQHDOLQJ .LQHWLFV RI *URXS,OO $FFHSWRUV LQ 2[LGL]HG 6LOLFRQ $SSO 3K\V YRO 1R -XQH

PAGE 172

%,2*5$3+,&$/ 6.(7&+ 0LFKDHO +DQ ZDV ERUQ LQ LQ 6LQJDSRUH +H UHFHLYHG D %6(( KLJK KRQRUVf DQG D 06(( GHJUHH IURP WKH 8QLYHUVLW\ RI )ORULGD LQ DQG LQ UHVSHFWLYHO\ ,Q KH ZRUNHG DW 7(&+ 6HPLFRQGXFWRU 6LQJDSRUHf D MRLQW YHQWXUH RI 7H[DV ,QVWUXPHQW 6LQJDSRUH (FRQRPLF 'HYHORSPHQW %RDUG &DQQRQ DQG +HZOHWW 3DFNDUG DV D 4XDOLW\ $VVXUDQFH (QJLQHHU +H WKHQ UHWXUQHG WR 8QLYHUVLW\ RI )ORULGD LQ WR FRQWLQXH KLV HGXFDWLRQ XQGHU WKH VXSHUYLVLRQ RI 3URI &KLK7DQJ 6DK DW WKH )ORULGD 6ROLG6WDWH (OHFWURQLFV /DERUDWRU\ +H ZLOO FRPSOHWH KLV 3K' GHJUHH LQ HOHFWULFDO HQJLQHHULQJ DW WKH 8QLYHUVLW\ RI )ORULGD LQ 'HFHPEHU

PAGE 173

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r 3HWHU =RU\ I\ \\ 3URIHVVRU RI (OHFWULFDO DP+MRPSXWHU (QJLQHHULQJ

PAGE 174

, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWDEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D WKHVLV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ n, $? A N A 8OULFK .XU]ZHJ 3URIHVVRU RI $HURVSDFH (QJLQHHULQJ 0HFKDQLFV DQG (QJLQHHULQJ 6FLHQFHV 7KLV WKHVLV ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH &ROOHJH RI (QJLQHHULQJ DQG WR WKH *UDGXDWH 6FKRRO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHPHQWV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 'HFHPEHU :LQIUHG 0 3KLOOLSV 'HDQ &ROOHJH RI (QJLQHHULQJ .DUHQ $ +ROEURRN 'HDQ *UDGXDWH 6FKRRO


8
Source (S)_
Device-
Gate (G)
IEEE488
Interface bus
Drain (D)
Instrument
Under-
Test
* *
(OUT)
J3ody/Base (B)
^Substrate (E^
Figure 2.1 Block diagram of the Stress-and-Measure SAM Station.
DEC
MicroVax II
Computer


55
surface layer underneath the gate oxide. The electron kinetic energy in the silicon
surface space-charge layer is determined by the reverse bias applied between the
n+source/n+drain and the p-well. The magnitude of the gate current can be adjusted by
the emitter-base forward bias voltage which determines the emitter current, and by the
collector-base reverse bias which controls the kinetic energy distribution of the electrons
at the Si/Si02 interface.
4.3 Results and Discussions
Figures 4.4 shows the IB-VGB and IG-VGB curves, measured in both the top-Emitter
and bottom-emitter configurations, and the Io-VDS curves measured before and after the
SHE stress. In Figure 4.4(a), the increase in the magnitude of the post-stressed IB peak
is attributed to an increase in the density of the stress-generated interface trap. The
VGB-it component in Eqn.(4.I) can then be determined using this stress-induced IB, AIB
= IB(post-stress) IB(pre-stress), using the following relations [74].
aib =
(qA^AS^) exp (qVBE/2kBT)
(4.6)
AS0 =
( 7T/2 ) (4.7)
AQit =
CqAVgb-it
(4.8)
=
qANIT
(4.9)
qADIT AEit
(4.10)
AS0 is the increase in surface recombination velocity due to the stress-generated
interface traps. AG is the gate area of the nMOST. n¡ is the intrinsic carrier
concentration. 0th is carrier thermal velocity. In eqns. (4.5) and (4.6), it is assumed that
the density-of-state (D!T) and the carrier capture cross sections (an = ap = o0) of the


30
p+Substrate
Figure 3.2 Cross-sectional view of n-channel MOST biased in the Drain-Emitter (DE)
DCIV configuration. Six base current recombination pathways are
depicted. White squares and hexagons in Si02 are oxide electron and hole
traps. White and black triangles are active (Si and SiO) and inactive-
hydrogenated (Si:H and SiO:H) interface traps. Square-grid-shaded areas
are space-charge regions of p/n and p+/p junctions in Si. Black and white
circles are electron and holes in Si. White squares in Si are electron-hole
recombination-generation centers. Adapted from Sah [63],


3
steps have been introduced to minimize the hot-carrier-induced device degradation [29-
32],
To evaluate the reliability of submicron MOST and its operation TTF without
actually operating the MOST continuously for 10 years, the semiconductor industry has
adopted a voltage-accelerated stress approach to extrapolate the TTF of a new or
production technology [33-37]. In this engineering approach, MOST fabricated from a
particular technology is stressed at higher voltages to accelerate its degradation rate.
Then the high-voltage accelerated-stress results are empirically extrapolated to
operating condition such as 5 V or 3.3 V. This engineering approach is highly empirical
and gives unreliable TTF extrapolation to lower voltages as it does not consider the
physical dependence of the device failure rate on the kinetic energy of the hot carriers
and the threshold energy for a particular denominate degradation mechanism and
pathway [38, 23]. In addition, an empirical failure model with fudging or adjustable
parameters does not provide any insight to the actual intricate degradation mechanisms
and pathways. Thus there is a need for a physics-based TTF extraction algorithm which
can provide fundamental physical parameters for evaluating a developing or production
technology and also for basic understanding of the basic physics of hot-carrier-induced
degradation in MOSTs.
Recently, Neugroschel and Sah [39-40] demonstrated a novel current-acceleration
(as opposed to voltage-acceleration) TTF extraction methodology to stress a Bipolar
Junction Transistor (BJT) at low emitter-base reverse-bias voltages. The current-
acceleration stress technique uses increased hot carrier density at low stress voltages to


17
conductive and transparent gate material or a special window on the substrate surface is
necessary. One key advantage of using SHEi is the independent control of three basic
stress parameters, the oxide electric field, the carrier kinetic energy and the hot carrier
current. The areally uniform electron injection into the gate oxide during SHEi also
simplifies the analysis.
Figure 2.6 shows the schematic SHEi setup for a nMOST with junction-well
isolation. During the injection stress, the gate of the nMOST is connected to a K486
picoammeter in series with a K230 voltage source. The K486 picoammeter measures
the gate current for fluence calculation. The K230 voltage source supplies a positive
voltage to the gate for maintaining a constant electric field across the gate oxide. The
source and drain are tied together and connected to another K230 voltage source where
a reverse bias is applied between the source/drain (collector) and the p-well (base) to
bias the vertical parasitic bipolar transistor into active region as well as to control the
amount of the band bending in the silicon surface space-charge layer. A forward-bias is
applied between the p-well (base) and the substrate (emitter) to supply the minority
carriers (electrons). A lk£2 resistor was connected in series with the emitter to limit the
emitter current. The emitter and collector currents were measured by two
programmable HP3478A ammeters. While keeping the reverse-bias between the p-well
(base) and source/drain (collector) constant, the forward bias across the p-well (base)
and substrate (emitter) was adjusted by a software feedback loop in the SAM control
program to maintain a constant gate injection current. The magnitude of the gate


19
injection current can be controlled by varying the forward-bias across the emitter-base
junction and the reverse-bias across the collector-base junction.
2.4.2 Channel Hot Electron Stress
The channel hot electron stress conditions in nMOST are classified based on the
magnitude of the applied VDS and VGS. Two commonly used CHE stress conditions are
(1) VDS < VGS or the "triode" range, and (2) VDS > VGS or drain current-saturation
range. When VDS < VGS, the transistor is operating in the linear range. The channel
electrons are accelerated by the large channel electric field in parallel with the Si02/Si
interface near the drain to gain more kinetic energy. Near the drain junction, the
electron trajectory may be deflected sideways towards the Si02/Si interface by phonon
scattering. If the kinetic energy of these impinging electrons is greater than the electron
barrier height at the Si02/Si interface (~3.12eV), they can be injected into the gate
oxide. This electron gate current was modeled using the Shockleys "lucky electron
model" [45] and expressed [46-48] as
1-G K IcheXP (2-1)
where Ich is the channel current, EM is the maximum electric field perpendicular to the
Si02/Si interface, and is the empirical electron scattering mean-free-path (=78A).
The electron barrier height, height given by [46]
where Eox is the oxide electric field near the drain junction. The last two terms takes
into account of the Schottky-barrier lowering and the possibility of tunneling through


129
(VD VDsat)/(1 V)
Figure 9.4 Plot of positive charge buildup as a function of the electron kinetic energy,
(VD VDsat), at a stress fluence of Ninj= 20x 1023 cm-2, from the data given
in Figure 9.3.


51
Source Gate Drain Base
n+
<
<
<
<
M
n+
p+base
contact
p-base well
n-epitaxial layer
n+ substrate
1
Substrate
Figure 4.1 Cross-sectional view of the BiMOS BJT-MOST transistor structure used.


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a thesis for the degree of Doctor of Philosophy.
'I A\ ^ (k ^
Ulrich Kurzweg
Professor of Aerospace Engineering,
Mechanics and Engineering Sciences
This thesis was submitted to the Graduate Faculty of the College of
Engineering and to the Graduate School and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy
December 1997
Winfred M. Phillips
Dean, College of Engineering
Karen A. Holbrook
Dean, Graduate School


135
SAM experiment on two 0.35 pm nMOST shown in Figure 9.9 stressed at two
traditional power supply voltages, VDS.stress= 5.0 V (Device-A) and 3.3 V (Device-B).
This short L= 0.35 pm MOST is fabricated from a 0.35 pm CMOS technology and has
a 6 nm thick gate oxide and a drawn channel width of 17.5 pm. For this L= 0.35 pm
nMOST, VDsat= 0.1679 V is determined from the conductance measurement.
Figure 9.9 plots the stressed-induced AIB_DE (left y-axis) versus VGB of Device-A
and Device-B stressed after 11.7 min and 494.8 min, respectively, and the unstressed
DE IB-VGB (right y-axis) characteristics of Device-A and Device-B (dotted line). The
stress-time difference was selected to make the stress drain fluence almost identical in
the two SAM tests. With reference to the unstressed IB peak at VGB= -0.45 V, the
stress-induced AIB_DE peaks for Device-A and Device-B both showed a negative shift
along the gate voltage axis. However, Device-A stressed at VDS_stress= 5.0 V was found
to have larger build-up of positive oxide charge +Q0T (AVGB 0.65 V = -AQot/Cox)
and interface traps NIT ( AIB_peak ~ 170 pA). Comparatively, Device-B stressed at
VDs-stress= 3.3 V exhibits a rather small +Qox build-up and NIT generation (AVGB -
-0.05 V and AIB_peak ~ 30 pA). This can be explained by considering the magnitude of
(VD VDsat) which is the acceleration voltage for the channel electron through the
drain/channel space-charge-layer at the Si02/Si interface. For Device-A, VD VDsat =
5.0 V-0.17 V = 4.83 V. For Device-B, VD VDsat = 3.3 V 0.17 V = 3.13 V. The
latter corresponds to the Auger threshold of 3.13 V. These two SAM results
complement the preceding long channel results that when VD VDsat < (0XhH EG)/q =
3.13 V, the +Q0t build-up from barrier-surmounting hot hole injection during CHE


161
[109] Yi Lu and Chih-Tang Sah, "Energy and Momentum Conservation during
Energetic Carrier Generation and Recombination in Silicon," Physical Review.
Vol.B52, No.8, pp.5657-5664, Aug. 15, 1995.
[110] Chih-Tang Sah, Fundamentals of Solid-State Electronics Solution Manual,
Table 915.2 on p. 121, World Scientific Publishing Co. 1996. New Jersey and
Singapore.
[111] J.F. Verwey, "On the emitter degradation by avalanche breakdown in planar
transistors," Solid-State Electron., vol.14, pp.775-782, Sep. 1971.
[112] T.H. Ning, P.W. Cook, R.H. Dennard, C.M. Osburn, S.E. Schuster, and H.N.
Yu, "lpm MOSFET VLSI Technology: Part IV Hot-electron design
constraint," IEEE Trans. Electron Devices. Vol.26, No.4, pp.346-353, Apr.
1979.
[113] T. Tsuchiya, T. Kobayashi, S. Nakajima, "Hot-Carrier-Injected Oxide Region
and Hot-electron Trapping as the Main Cause in Si nMOSFET Degradation,"
IEEE Trans. Electron Devices, Vol.34, No.2, pp.386-391, Feb. 1987.
[114] Chih-Tang Sah, "VLSI device reliability modeling," Proc. 1987 International
Symposium on VLSI Technology, Systems and Applications, Taipei, May 13-
15, 1987, pp.153-162.
[115] C.J. Varker, D. Pettengill, W. Shiau, and B. Reuss, "Hot carrier hE degradation
in BiCMOS transistors," Proc. 30th Int. Reliability Phys. Symp., pp.58-62,
1992.
[116] I.C. Kizilyalli and J.D. Bude, "Degradation of gain in bipolar transistor," IEEE
Trans. Electron Devices, Vol.41, No.7, pp. 1083-1091, Jul. 1994.
[117] Chih-Tang Sah, Fundamental of Solid-State Electronic Study Guide, Fig.
B2.3 to B2.6 on pp.404-412, World Scientific (Singapore), 1993.
[118] Private Communication of C.T. Sah to M. Han on June 18, 1997 and A.
Neugroschel on June 17, 1997.
[119] Karl R. Hofman, Christoph Werner, Werner Weber, and Gerhard Dorda, "Hot
Electron and Hole Emission Effects in Short n-Channel MOSFETs," IEEE
Trans. Electron Devices, Vol.32, No.3, pp.691-699, Mar 1985.
[120] P. Heremans, H.E. Maes and N. Saks, "Evaluation of Hot Carrier Degradation
of N-Channel MOSFETs with the Charge Pumping Technique," IEEE
Electron Device Letters, Vol. EDL-7, No.7, pp.428-430, Jul 1986.


38
Space-charge layer
n+ Poly Si
Si02
p-Si
n+ Si
Gate
Oxide
Base
Emitter
Figure 3.5 Cross-sectional view of the transition energy band diagram.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EMYONRU6A_4B1Z7U INGEST_TIME 2014-11-17T21:18:19Z PACKAGE AA00024479_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


35
Si02
interface traps in this
thin interfacial layer
Figure 3.4 Cross-sectional view of a thin interfacial layer at the Si02/Si interface.


15
procedure. Details of DCIV measurement configurations and its basic principles will be
elaborated in Chapter 3. Figure 2.5 shows the schematic setup for Drain-Emitter DCIV
measurement. A K230 voltage source (constant VEB) is connected to the drain to
forward bias the drain n+/p junction. Another K230 voltage source (VGB) is connected
to the gate to modulate the Si surface from accumulation to inversion. A K485
picoammeter is connected to the p-well to measure the body current at each VGB step.
The source is grounded during the entire measurement.
2.3.3 Gated IB-VBB Measurement
The IB-VBE measurement is identical to the standard method used for bipolar
transistors characterization. The only difference here is that a constant d.c. voltage is
applied to the gate of the MOST during the IB-VBE measurement. In the case of a
standard 3-terminal BJT structure, there is no gate terminal available. The IB-VBE
measurement uses exactly the same DCIV setup shown in Figure 2.5. During the
measurement, the emitter-base (for example the n+drain/p-body-well junction is
forward-biased and the VBD is increased with small incremental steps while a constant
voltage is maintained at the gate. The base (or body ) current is monitored using the
K485 picoammeter.
2.4 Substrate and Channel Hot Electron Stress
In silicon, thermal electrons at the conduction band edge can gain kinetic energy
by tranversing through a potential drop and becomes energetic and hot. These hot
electrons are characterized by an effective temperature, Teff = (2/3)(Kinetic
Energy)/kB, which is much greater than the lattice temperature, TLattice. For example, if


9
leakage-low noise 50Q cables to the instrumentation rack. Data acquisition and
instrumentation control are automatically controlled by the SAM programs in the
MicroVax-II computer through a IEEE-488 interface bus. All the SAM control
programs were coded in the VAX-FORTRAN language. A flow chart of the SAM
control program is summarized in Figure 2.2.
On the instrumentation rack, a 6-1/2 digits high resolution HP34401 digital
voltmeter is used to measure the voltages supplied by three programmable voltage
sources: two Keithley K230 and one Hewlet-Packard HP6106 power supplies. The
Keithley picometers (K485 & K486) and HP3478 multimeters are used to measure the
d.c. currents during stress and measurement cycles (see Figure 2.3). The measurement
and stress configurations are multiplexed by a HP3488 scanner and a custom-built Reed
switching relays box. The HP3488 scanner contains two relays cards: (1) a 10-channel
array of single-pole/single-throw relays controls the Reed relay box, and (2) a 10-
channel array of double-pole/single-throw relays is multiplexed together with the
HP34401 voltmeter to measure the terminal voltages applied on DUT.
2.3 Electrical Measurement
2.3.1 Drain-Current Measurement
Two drain current measurement configurations are implemented by the SAM
station: (1) linear Io-VGS, and (2) saturation Ipsat'^DS- The difference between the
linear and saturation drain current measurement depends on how the drain and gate are
connected electrically. Figure 2.3 and Figure 2.4 illustrate the schematic of the SAM
station setup for linear and saturation ID measurement, respectively. During the linear


-12
113
Figure 8.7 DE-DCIV Degradation characteristics of 0.35 |im nMOST under CHE
stress. CHE stress condition: VDS= 3.65V, VGS= 0.5V, VSB= 0.0V.


131
Figure 9.6 Kinetics of positive oxide charge buildup versus stress fluence, Ninj, in 20
pm x 20 pm nMOSTs stressed at VDS= 5.0 V, 5.2 V, 5.4 V, 5.6 V, 5.8 V,
6.0 V, 6.2 V, 6.5 V, and 7.0 V, and the same VGS= 0.8V


107
eV ( pMOST. The bond-breaking can be compared with the handbook Si:H bond energy of
3.101 eV. The VDS-independent bond-breaking contribution from channel hot electrons
in nMOST and hot holes in pMOST, appears to be absent, just like the BJT case, giving
one experimental evidence that the hot carriers injected from the drain edge are
responsible for generating interface trap. This observation was first ascertained in [42],
Hence, for nMOST, the channel-electron-impact-generated holes, injected from the n-i-
drain region into the p-body, break the Si:H and SiO:H bonds in the drain space-charge
region to generate new interface traps. For pMOST, these Si:H and SiO:H bonds are
broken by the channel-hole-impact-generated electron.
8.4.2 Short Channel CACHE
Following the CACHC procedure established for the long channel MOSTs in the
preceding section, state-of-the-art quarter-micron production nMOST was next used in
the CACHE SAM experiment. The quarter-micron nMOST is designed for 2.5 V
circuit operation and has 6nm oxide and drawn channel length and width of 0.35 Jim
and 17.5 p.m. The effective or metallurgical channel length of this submicron MOST is
-0.25 (im due to the lateral diffusion of the donor impurities during the n+source and
n+drain junction formation. Since the production nMOST is fabricated using CMOS
technology with p-epilayer on p+ base substrate, the n+source/p-Body junction is the
only p/n junction that can be forward-biased during CACHE stress. A forward bias of
VBS= 0.8 V was used. The CHE-induced degradation of the nMOST is monitored using
"reverse" bias configuration which gives ISsat, IS-VG and DE-DCIV since the damages is


123
9.4. Results and Discussions
9.4.1 Drain-Emitter Characteristic
A typical family of DE-DCIV curves at VDS_stress= 7.0 V and VGS,stress= 0.9 V and
stress times of 0, 0.1, 1, 11,26 ksec from a SAM run taken on a 100 |im/100 p.m
nMOST is shown in Figure 9.2. The nearly stationary peak at VG= -0.3 V was
identified by its exp(qVBD/kBT) dependence as recombination of n+drain-junction-
injected electrons with the p-body holes at the fabrication-residual interface located in
the quasi-neutral channel region out of the drain-junction space-charge region [59]. Its
growth rate is very low because there are few hot electrons and holes during stress in the
inverted channel outside of the drain-junction space-charge-region.
The fast growing secondary peak, VG= -2.25 V after 26 ksec stress, was identified
by its exp(qVDB/2kBT) dependence (see inset of Figure 9.2) as electron-hole
recombination at the stress generated interface traps covering the space-charge-region of
the n+drain/p-channel junction [59, 60, 97, 100, 111]. The rapid rise of this peak was
shown in Chapter 7 to come from interface trap generation by the secondary hot holes
from interband-impact generation of electron-hole pairs initiated by the primary channel
hot electrons. The silicon and oxygen dangling bond interface traps are generated by
hole capture or impact which releases the hydrogen of the Si:H and SiO:H near the drain
junction. The spatial location of the interface trap can be estimated at flat-band
condition using the one-dimensional n+/p junction space-charge-layer thickness
formula.


18
K230
Figure 2.6 SAM schematic setup for BiMOS-SHE injection stress.


158
[77] Andreas Schwerin, Wilfried Hansch, and Werner Weber, "The Relationship
Between Oxide Charge and Device Degradation: A Comparative Study of n-
and p- Channel MOSFETs," IEEE Trans. Electron Devices, vol.24, no. 12,
Dec. 1987.
[78] T. Mizuno, T. Kovori, Y. Saitoh, S. Sawada, and T. Tanaka, "Gate-Fringing
Field Effects on High Performance in High Dielectric LDD Spacer
MOSFETs," IEEE Trans. Electron Devices, vol.39, no.4, pp.982-989, Apr.
1992.
[79] T. Wang, C. Huang, P.C. Chou, Steve S.-S. Chung, and T.-E. Chang, "Effects
of Hot Carrier Induced Interface State Generation in Submicron LDD
MOSFETs," IEEE Trans. Electron Devices, vol.41, no.9, pp.1618-1622, Sep.
1994.
[80] S. Shabde, A. Bhattacharyya, R.S. Kao, and R.S. Muller, "Analysis of
MOSFET Degradation Due to Hot-Electron Stress in terms of Interface-State
and Fixed-Charge Generation," Solid-State Electron., vol.31, no.l 1, pp.1603-
1610, May 1988.
[81] Y.S. Jean and C.Y. Wu, "The Threshold-Voltage Model of MOSFET Devices
with Localized Interface Charge," IEEE Trans. Electron Devices, vol.44, no.3,
pp.441-447, Mar. 1997.
[82] P. Roblin, A. Samman, and S. Bibyk, "Simulation of Hot-Electron Trapping
and Aging of nMOSFETs," IEEE Trans. Electron Devices, vol.35, no. 12,
pp.2229-2237, Dec. 1988.
[83] H.S. Haddara and S. Critoloveanu, "Parameter Extraction Method For
Inhomogeneous MOSFETs Locally Damaged by Hot Carrier Injection," Solid-
State Electron., Vol.31, No. 11, pp.1573-1581, May 1988.
[84] P. Speckbacher, A. Asenov, M. Bollu, F. Koch and W. Weber, "Hot-Carrier-
Induced Deep-level Defects from Gated-Diode Measurements on MOSFETs,"
IEEE Electron Device Letters, vol.ll, no.2, pp.95-97, Feb. 1990.
[85] Mario G. Ancona, Nelson S. Saks and Daniel McCarthy, "Lateral Distribution
of Hot-Carrier-Induced Interface Traps in MOSFETs," IEEE Trans. Electron
Devices, vol.35, no. 12, Dec. 1988.
[86] Paul Heremans, John Witters, Guido Groeseneken and Herman E. Maes,
"Analysis of the Charge Pumping Technique and Its Applications for the
Evaluation of MOSFET Degradation," IEEE Trans. Electron Devices, vol.36,
no.7, pp. 1318-1335, Jul. 1989.


34
hole through the interface trap which has a discrete energy level Ej in the silicon energy
gap. The following presentation is based on the lectures given by Sah [69]. Consider
an infinitesimal interfacial layer of thickness 8x at the Si02/Si interface where the
electron-hole recombination occurs (Figure 3.4), the steady-state electron-hole
recombination rate Rss (cm-2s-1) at the interface traps, in the shaded area region, is
pss ~
cnscpsNSPS enseps
cnsNS + ens + cpsPS + ep
cnscps^SPS enseps
CnsNS + ens + CpsPS + ep
Nit(x) 6x dx
N
IT
(3.3)
(3.4)
where cns and cps are the electron and hole capture rate coefficients (cm3sec-1), ens and
eps are emission rate coefficients of the trapped electron and hole (sec-1), Ns and Ps are
the concentrations of electron and holes at the interface (cm-3), and NIT is the areal
density of interface trap (cm-2). At and near thermal equilibrium (i.e. neglect electric
field or hot carrier effect) ens can be related to cns by
ni exp(
Similarly for eps and cDS,
ps
'ps
n exp(
Et Ej
kBT
E-j- Et
) = n-i
(3.5)
'ps
kBT
) = Pi
(3.6)
In Eqns.(3.5) and (3.6), n¡ is the intrinsic carrier concentration of silicon, Ej is the
intrinsic energy level, kB is the Boltzman constant and T is the ambient temperature.


72
the experimentally observed n=1.7 ideality factor obtained by fitting AIB_DE-VBE to a
single exponential function [AIB=AIBOexp(qVBE/nkBT)].
Figures 5.4(a) and 5.4(b) show the post- and pre-stressed Source-Emitter IB-VCB
experimental results, AIB-VCB and -[9(AIB)/dVCB], respectively. In Figure 5.4(a), the
unstressed IB is essentially flat for -0.2V < VCB < 2.5V which suggests low and
uniform interface trap density. This is intuitively expected for an unstressed MOST
fabricated by the state-of-the-art MOS technology. The sharp rise in IB for VCB <
-0.4V is due to the additional base current contribution by the slightly forward-biased
drain-body junction. This excess base current has an exponential dependence on the
magnitude of VCB which will overwhelms the background base current. The slope of the
post-stressed IB data is observed to change at different value of VCB. This suggests a
non-uniform distribution of stress-generated NIT. The downward trend of both the post-
and pre-stressed base current in Figure 5.4(a) when VCB > 3.0V indicates hole current
from interband impact generation by the injected electrons which are accelerated by the
high electric field in the revere-biased (~3V) drain junction space charge-layer. Thus,
this limits the range of VCB that can be applied to extend the drain/body space-charge-
layer out into the channel without causing any additional stress on the device during
profiling measurement. To extract the stress-generated N1T from the measured post- and
pre-stressed IB-VCB, we perform the following analysis on the data. The difference
result is shown in Figure 5.4(b).
AIb s [ IB (VBE=0.4V, tstr=12 Omins) IB (VBE=0.4V, tstr=0) ] -
[IB (VBE=0.0V, tstr=120mins) -IB (VBE=0.0V, tstr=0) ] (5.9)


11 SUMMARIES AND CONCLUSIONS
148
REFERENCES 151
BIOGRAPHICAL SKETCH 164
v


50
injection (SHEi) to uniformly generate QIT and Qqt- The transistor has a drawn
channel length of 1.6|im and a width of 100|im. The gate oxide, thermally grown at
850C in a dry oxidation ambient, is 150A thick. A cross-sectional view of the nMOST
is shown in Figure 4.1. The stress-generated QIT and QOT are monitored by the top- and
bottom-emitter DCIV, and the conventional ID-VG measurements. The top- and
bottom-emitter measurement configurations are shown in Figure 4.2.
Figure 4.3 shows the energy band bending of the gate oxide and the Si substrate in
the nMOST during SHE injection. A positive voltage (VG=7.5V) is applied between
the gate and the source and drain junctions which are connected together to invert the
channel underneath the gate oxide. This inverted n-channel, together with the n+ source
and drain junctions, serves as the collector for the vertical n+/p/n-channel bipolar
transistor. The p-well is the base and the n-/n+ epitaxial substrate serves as the electron
emitter. The emitter/base junction is forward-biased to inject electrons into the p-well
base. These injected electrons diffuse upwards through the p-well and are accelerated
towards the Si/Si02 interface by the electric field in the surface space-charge layer of
the reverse-biased collector/base n-channel/p-well junction. The reverse bias, VCB,
provides the accelerating field for these electrons. When the channel is strongly
inverted, VCB will be the potential difference between the hole quasi-Fermi potential ,
Vp=FP/q, in the p-well and the electron quasi-Fermi potential, VN=FN/(-q), at the
Si/Si02 interface. The accelerating electrons may acquire sufficient energy in the Si
space-charge layer to surmount the Si/Si02 barrier (3.12eV). Hence, the minimum VCB
that must be applied to accelerate the injected electrons into the gate oxide is about


monitored transistor degradation parameters: drain-saturation current, gate threshold
voltage, and subthreshold voltage swing (AIDsat, AVGT and AS).
A physics-based Time-to-Failure extraction algorithm is developed using the new
current-accelerated channel-hot-carrier stress (CACHC) method. Current-acceleration
of the transistors degradation rate is obtained by forward-biasing a p/n junction to
increase the channel hot carrier current. The applied stress voltages can be
independently set to control the kinetic energy of the channel hot carriers. Stress-time
reductions by about 1-2 orders of magnitude are demonstrated. Correlation with Sahs
bond-breaking interface generation model showed that interfacial silicon-hydrogen
bonds are broken by hot holes in n-channel MOST and by hot electrons in p-channel
MOST.
Positive charging of oxide trap by barrier-surmounting hot hole during channel hot
electron stress is investigated using the DCIV technique. Two pathways are
experimentally observed: (1) hot-electron/hot-hole interband impact pathways in high-
voltage long-channel nMOST with electron kinetic energy threshold of 5.12 eV, and (2)
hot-electron/thermal-hole Auger recombination pathway in low-voltage short-channel
nMOST with electron kinetic energy threshold of 3.13 eV.
Interface trap generation in p-channel MOST during channel hot hole stress is
investigated. Reduction of interface trap density is observed for the first time.
Hydrogen released from the hydrogenated boron in the p+drain region by the channel
hot hole is the proposed mechanism. The released hydrogen diffuses from the p+drain
vii


21
operation and cause a snap-back breakdown and CMOS latchup. Due to the 2-
dimensional electric field distribution in the drain space-charge layer, empirical models
with adjustable parameters have been used to model Ix by relating the peak electric field
(Em) at the drain to the drain voltage (VD) [33, 35]
Ix = A Ich exp (B/Em) (2.3)
where A and B are empirical fitting parameters. The peak well current was assumed to
have an exponential dependent on the maximum electric field at the drain edge. An
estimate of the maximum channel electric field was given by [35],
vD vDsat
Em = (2.4)
^ Y dep
where VDsat is the drain saturation voltage, L is the channel length, and ydep is channel
depletion point. Eqn.(2.4) assumes that the maximum longitudinal electric field in the
silicon surface layer between the source and drain occurs between the depletion point,
y=ydep, and the drain metallurgical junction, y=L. The distance from the channel
depletion point to the drain metallurgical junction has been empirically modeled by
Chan-Po-Hu [56],
L ydep = 0,2.xox1/2.Xj1/3 (2.5)
where x0X is the gate oxide thickness and Xj is the drain junction depth.
At large VDS and VGS= VGT energetic impact-generated electrons and holes can be
injected sideway into the gate oxide [57-58]. Between the depletion point and the
source, the surface and oxide electric field attracts the channel electrons to the Si02/Si
interface. The surface and oxide electric field is directed in the opposite direction


149
monitor MOS transistor degradation. A turn-key DCIV software was developed and
delivered to members of the Semiconductor Research Corporation.
A physic-based Time-to-Failure extraction algorithm was demonstrated using the
current-acceleration CHC stress technique. The current acceleration method uses a
forward-biased p/n junction to increase the channel hot current to enhance the
transistors degradation rate. The advantage of this current acceleration over the
traditional voltage-acceleration CHC stress method is that the kinetic energy of the
channel hot carriers can be maintained close to operating condition such that the same
degradation mechanisms responsible for transistor degradation during operation can be
simulated by the current acceleration CHC stress. Stress-time reduction by about 1-2
orders of magnitude is demonstrated to determine the maximum operation voltage for a
ten-year time-to-failure in n- and p-channel MOST. This technique will be attractive
for advanced deep-submicron low-voltage MOST development and evaluation. The
CACHC results were also correlated with Sahs bond-breaking interface trap generation
theory and showed that interface traps generated during CHC stress has a 3.06 eV
threshold energy. Silicon-hydrogen (Si:H) bonds at the Si02/Si interface are concluded
to be broken by channel hot holes in nMOST and by channel hot electrons in pMOST
during high-voltage stress. A 1-V threshold interface trap generation via Auger-
recombination is identified to be a dominant degradation mechanism for low-voltage
(3.3V and below) MOST.
During CHE stress, oxide hole traps over the drain junction space-charge layer are
charged by capturing holes injected into the oxide over the Si02/Si hole barrier. Four


144
10
-9 -
CD
<
n=2
t0-5
i a
a-'
i rn i r
961224MH
saturates
--A-Alii A A
iO"10h
- t1V
A
DCIV Vdb=0.5V
vCB=ov
A DE-DCIV Vgb=2.0V
10
,-11
(a)
J i''l I L-LlL I I L-lJ I L_uJ I L.
Figure 10.3 Interface trap generation kinetics obtained by regular and fast DCIV
sampling of stressed pMOSTs by channel hot holes at VDS= -5.25 V, VGS=
-1.0 V, and Vsx= 0 V. (a) Fast DCIV sampling at VDB= 0.5 V and VGB=
2.0 V. (b) Regular DCIV at VEB= 0.6 V at VGB= 2.0, 0.2 and -0.2 V.


14
K230
Figure 2.5
SAM schematic setup for DCIV measurement showing the Drain-emitter
DCIV configuration.


82
The stress-induced widening of the subthreshold swing, AS, is also shown in
Figure 6.4. It increases initially as AS AIB, then rises faster, and finally approaches
again AS oc AIB. This is consistent with the simple subthreshold slope expression,
S = 2.303 (kT/q) [l+(Cit+Cs)/Cox] (6.3)
where Cit = qD[T and NIT = j D IT(ET)fT(ET)dET [100], which indicates an initial and
final linear dependence, AS ACit ANIT = AIB, while the superlinear rise of AS in the
mid-range of stress is caused by the faster change of the surface potential with AVGT
relative to the energy distribution of DjjfEj).
In summary, an experimental proof is given of the theoretically anticipated linear
dependence of AIB_peak on the traditionally monitored transistor reliability
characteristics, AID_sat and AVGT. This correlation provides the basis for using the
DCIV method to monitor MOS transistor degradation.


CHAPTER 9
HOT HOLE INJECTION INTO Si02 IN N-CHANNEL MOS TRANSISTOR
DURING CHANNEL-HOT-ELECTRON STRESS
9.1 Introduction
A decrease in threshold voltage [113, 119] and an increase in transconductance
[120-123] were observed when short n-channel metal-oxide-silicon transistors (MOST)
are stressed under high VD and low VG ~ VGT conditions. Positive charge created in the
Si02 near the drain region during stress was assumed to account for the threshold
voltage decrease and transconductance increase. Several workers speculated that the
positive charge is due to the positive charging of the oxide traps by the hot holes
generated from the interband impact process near the high-field drain junction [55, 57-
58, 113, 120], For the nMOST biased in the high-VD and low-VG condition, the surface
energy band bending near the drain region will favor the hot hole injection into the gate
oxide [50, 119, 124]. Hofmann et al [119] employed a 2-D effective hole temperature
model in MINIMOS 2 to simulate hot hole injection into gate oxide and reported that
the hole injection can occur under high VD and low VG conditions. This was later
verified by Nissan-Cohan [57] and Saks et al [58] who employed a floating-gate
technique to measure the ultra-low gate current (in the order of 0.01 fA). They
concluded that at high-VD and low-VG (~VGT) stress condition, the ultra-low positive
gate current measured in their experiment is due to hot-hole surmounting the Si02/Si
energy barrier ((|)H_h= 4-25 eV) and injected into the Si02. Injected holes can be
117


63
^B-IC
q W
Lys
Rss dY-
yD
(5.1)
In Figure 5.1, the drain-body metallurgical n+/p junction is referenced as the origin of
the y-direction. Thus, yD is the drain surface-charge layer thickness, ys is the source
surface space-charge layer thickness, and L is the channel length.
In Eqn.(5.1), the surface recombination rate Rss can be independently controlled
by VGB and VBE. VGB controls the surface band-bending condition and can be related
to surface band-bending potential \|/s using Eqn.(3.32). VBE determines the separation
between the quasi-Fermi potentials for electron FN and hole Fp and is related by VBE =
(Fn Fp)/q. As shown in Eqns.(3.9) and (3.12), the surface band-bending and quasi-
Fermi level splitting control the surface electron and hole concentrations. Expressing
Rss = NIT(y)R s.eff where NIT is the areal density of the interface traps and Rs.eff is the
effective recombination rate, IB_1C can be further expressed as
rLys
IB-ic = q w
NIT(y) Rs_eff dy.
(5.2)
yD
It is reasonable to assume that the surface band bending (\|/s) and the quasi-Fermi
potentials (FN and Fp) are position-independent for small channel current and long
minority carrier diffusion length (LDiff > L). Then Rs.eff can be treated as a constant and
independent of position and Eqn.(5.2) becomes
I
B-1C
q
W R
S-eff
L-ys
Nit (y) dy.
yD
(5.3)


66
of the spatial distribution of the interface trap in the region near the drains metallurgical
p/n transition boundary.
To obtain a qualitative value for NIT, a model for Rss needs to be assumed.
Analytical models for Rss can be a single energy-level QIT [59, 97] or a distributed DIT
that is uniformly distributed in energy near the middle of the silicon energy gap [66,
73]. The conversion of VDX to y-position to give the y-dependence of N1T can be
obtained using a MOST device simulator with known doping profile and devices
structure. However, an analytical abrupt n+/p junction approximation [98], given
below, can be used as a first order estimation to relate the drain surface space-charge
layer thickness to the drain voltage.
= t (VDX_bi+VDSl/rs) ]1/2 (5.8)
where eSi is the silicon dielectric constant, VDX_bi = (kBT/q)loge(NDD+NAA/n¡2) is the
built-in potential of the drain n+/p junction at the interface. The silicon surface energy
band bending in the x-direction at the drain, \}/s, depends on the applied gate voltage and
can be determined from the one-dimensional electrostatic formula of a MOS capacitor.
Ndd+ is the donor concentration in the heavily-doped drain and NAA is the acceptor
concentration in the p-base. Note that in the case of MOST with a lightly-doped-drain
structure, the yo(^Dx) estimated using Eqn.(5.8) will be larger than the actual value as
the voltage drop across the lowly-doped drain region is not considered by this simple
approximation.
Based on Eqn.(5.7), qualitative interpretations on the spatial distribution of the
interface traps can be deduced. For example, if there is a non-uniform distribution of


16
KE= 1 eV, Teff = (2/3).300K(leV/0.025eV) = 8,000 K Tlattice= 300 K at room
temperature. In nMOST, there are two important hot-electron effects: (1) the substrate
hot-electron (SHE) effect where the primary source of hot electrons is due to the current
from the substrate, and (2) the channel hot-electron (CHE) effect where the primary
source of hot electrons is due to the surface channel current. To investigate these two
hot-electron effects in silicon MOST, the stress mode implemented on the SAM station
is carefully designed by considering the desired stress conditions and configurations.
These will be described in the following two sections.
2.4.1 Substrate Hot Electron Stress
In this work, a substrate hot electron injection (SHEi) technique is used to
uniformly inject electrons from the silicon substrate into the gate oxide of the MOST.
This injection technique employs a forward-biased vertical bipolar transistor [16] and
has been previously used to employed to study the basic degradation mechanisms of the
Si02 gate oxide and its fundamental electrical properties [20, 22, 24, 43], Optical hot
electron injection (OHEi) [44] is a variation of SHEi. The major difference between
OHEi and BiMOS-SHEi is the source of minority carriers (electrons) in the p-type
silicon body of the nMOST. For OHEi, the hot electrons comes from interband optical
generation of electron-hole pair in the surface space-charge-region in the p-silicon
substrate underneath the gate oxide. For BiMOS-SHEi, the source of hot electrons is
the minority carriers (electrons) injected by the forward-biased n+/p emitter-base
junction. BiMOS-SHEi is employed in this work because of its easy and
straightforward implementation. For OHEi, a special test structure with thin layer of


109
R_VBB+Esi-gap/q] > (8.4)
The operation TTF, computed using Eqn. (8.2), is shown on the left-axis of Figure
8.9(b). Using the failure criteria of 5% IDsat reduction or lOmV AVGT increase, the
theoretically extrapolated maximum VDD voltage for this 2.5 V CMOS technology is
2.48V for a 10-year operation lifetime. For the traditional voltage-acceleration
extrapolation method shown in Figure 8.10, to give a 10-year operation lifetime, the
maximum operating voltage is only 2.09V. The traditional voltage acceleration method
is clearly in error as it empirically assumes that the degradation of the MOST is
exponentially related to the reciprocal of the drain-source voltage [33-37] and,
disregarding any threshold kinetic energy, using an operation TTF given by
TTF = Aexp(B/VDS) (8.5)
where A and B are empirical fitting constants.
A closer examination of Eqn. (8.4) gives a most important result (derived by Sah)
[118]. The hole injection from the drain edge due to this Auger-recombination pathway
[117] has the lowest threshold for interface trap generation. This implies that interface
trap generation will remain a dominant degradation process for low-voltage MOST as
long as the supply voltage is greater than IV. This is illustrated by the following
relation.
VD VDsat (VBB
(8.6)
(3 1)/2
IV


91
details of the nMOST and pMOST SAM data will be described in Chapter 8. The two
sets of npnBJT data are obtained from emitter-base reverse-bias stress experiment. The
npnBJTs were fabricated from a submicron BiCMOS technology with 7xl017 cm-3 base
doping concentration and 10|imx0.6|im emitter contact area. During the emitter-base
stress experiment, the emitter-base junction of the npnBJTs was reverse-biased at VR
while the collector-base junction was either open-circuited (OC) or forward-biased
(FC). A 10% change in the common emitter d.c. beta gain was used as a failure criteria
in the TTF SAM experiments.
Figure 7.2 plots the experimental and theoretical interface trap generation
efficiency for the nMOST, pMOST and npnBJTs versus the applied stress voltage. The
interface trap generation efficiency was obtained by taking the inverse of the TTF SAM
data and translating the efficiency data vertically by normalizing for the difference in
stress current density and contact area in each SAM experiment. The applied stress
voltage is the magnitude of the drain voltage for the nMOST and pMOST data and the
magnitude of reverse-biased emitter-base voltage for the npnBJTs. The nMOST and
pMOST experimental data were least-square-fitted using Eqn.(7.12) where VB1=0 and
VR=VD-VDsat. The LSF results give bond-breaking energies of qVBB= 3.087 eV and
standard deviation of 1.12 % for nMOST, and 3.040 eV and 1.67 % for pMOST. The
HFHi npnBJT data (inverted triangle in Figure 7.2) was also least-square-fitted using
Eqn.(7.12). The LSF results gives a bond-breaking energy of qVBB= 1.876 eV and a
standard deviation of 1.51%. A reasonable built-in potential of VBI= 1.11 V was also
obtained from the LSF, reflecting the highly doped base region in advanced submicron


25
K230
Figure 2.9 SAM schematic setup for current-accelerated CHE stress using the
forward-biased substrate/body junction.


112
VQD/(1 V)
Figure 8.6 Degradation characteristics of 0.35 fim nMOST under CHE stress. ID-VG
is measured in the "reverse" configuration to monitor the degradation near
the drain junction. Measured at VDS= 0.5 V. CHE stress condition: VDS=
3.65 V, VGS= 0.5 V, VSB= 0.0 V.


145
The decreasing peak starting at VGB= -0.5 V, which is plotted at VGB= -0.2 V in
Figure 10.3, follows a square-root stress-time dependence, suggesting that the hydrogen
diffusion is the rate-limiting step in the hydrogenation of the interface traps in the mid
channel region. This is consistent with the diffusion delay of the hydrogen released by
channel hot holes from hydrogenated boron, Si:BHSi, in the p+drain. Boron
hydrogenation was first identified and verified by Sah-Sun-Tzouo in a series of MOS
capacitor avalanche experiments [132-134]. Three additional observations supported
this model: (1) the growth rate of the decreasing peak at VGB= -0.5 V is much lower in
long-channel pMOSTs than in short channels, (2) a high density of hydrogenated boron
is available in the heavily-doped p+drain [a 1% boron hydrogenation in the heavily-
doped p+drain (~1020 Boron/cm2) contains ~1018 boron-hydrogen bonds], and (3) the
decreasing peak is barely noticeable in nMOSTs which suggested that the source of the
hydrogen was from the hydrogenated boron-acceptor in nMOSTs p-well which has a
much lower boron concentration than the p+drain of the pMOST.
Figure 10.4 illustrates the hydrogen diffusion pathways for the CHH-stressed
pMOST. Hydrogen released from the broken Si and SiO. bonds and dehydrogenated
boron acceptors Si:B can either diffuse towards the p+poly-Si, along the Si02/n-Si
channel or in the n-Si substrate. However, hydrogen diffusion towards the p+poly-Si
does not participate in the hydrogenation process at the Si02/Si interface though it
occurs at a faster rate than its lateral diffusion towards the mid-channel region, as shown
by the following estimation using LDiff = V(Dt).


125
YSC = [2£siVPN/qNAA)1/2 (9.5)
= 36[(Vpn/1 V)(1016/Naa)]1/2
With Naa= 5x1016 Boron/cm3, then during DE-DCIV measurements,
VpN = VBI-VBD (9.6)
= (0.95-0.4) V
= 0.55 V
Substituting VPN= 0.5 V into (9.5), the drain space-charge-layer thickness during DE-
DCIV measurement is estimated to be
YSC-me,s = [2esi.0.5V/qNAA)' (9.7)
= 36[(0.55/lV).(1016/5xl016)]1/2nm
= 12 nm
During CHE stress at VDS= 7.0 V and VGS= 0.9 V,
VPN = VD-VDsat (9.8)
= 7.0 V-0.33 V
= 6.67 V
then, the drain space-charge-layer thickness is estimated to be
Ysc-,, = [2esi.0.5V/qNAA)1'2 (9.9)
= 36[(6.67/l V).(1016/5xl016)]1/2nm
= 41 nm
Based on these estimations, the fast-growing interface traps are located in the channel
less than 12 nm from the n+/p-surface boundary.


ACKNOWLEDGEMENTS
I wish to express my gratitude to Professor Chih-Tang Sah for all his wisdom,
guidance, patience and time throughout the course of my graduate research at the
University of Florida. I would also like to express my thanks to Professors Arnost
Neugroschel, Toshikazu Nishida, Peter Zory and Ulrich Kurzweg for serving on my
supervisory committee.
I wish to thank my past and present colleagues, Dr. Scott Thompson, Dr. Yi Lu,
Dr. Jack Kavalieros, Dr. Michael Carroll and Steven Walstra, for helpful discussions.
Special thanks to Steve Walstra for his friendship and the many interesting and insightful
discussions and debates. Assistance by Derek Martin and Jin Cai is also appreciated.
Financial support from Semiconductor Research Corporation is gratefully acknowledged.
I am indebted to my family for their persistent caring and support throughout
this endeavor.
11


CHAPTER 5
PROFILING OF INTERFACE TRAPS
BY THE DCIV METHOD
5.1 Introduction
Due to the localized nature of channel-hot-carrier induced degradation, multi
dimensional transistor simulators such as PISCES or MINIMOS are usually employed
to analyze the effects of channel-hot-carrier induced degradation on the electrical
performance of the transistors [50, 76-83]. However, lack of reliable experimental data
has often required using assumed spatial distribution of interface and oxide in transistor
simulations. Hence, there is a need for a reliable characterization technique to profile
the spatial distribution of stress-generated interface and oxide traps where the
experimentally-determined interface and oxide trap profile can be fed into the
simulators to analyze the electrical characteristics of a degraded transistor.
Two experimental techniques were recently proposed to characterize the spatial
distribution of stress-generated interface traps (QIT) near the drain junction of a n-
MOST [84-88]. Speckbacher et al [84] used the d.c. surface generation current
technique to profile the spatial distribution of QIT. In this technique, the weak reverse
biased drain junction current was measured in a gated diode configuration and the gate
bias is varied from accumulation to inversion to vary the surface space-charge layer near
the drain junction. Due to the small reverse bias applied on the drain, the profiling
61


41
£si(dE/dx) = -£si (d2V/dx2)
= q ( P-N-Pb + Nb) (3.17)
where £Si is the silicon dielectric constant, E is the electric field strength into the bulk
Si, P is the hole carrier concentration, N is the electron carrier concentration, PB is the
bulk hole concentration, and NB is the bulk electron concentration. The boundary
conditions at the Si02/Si interface and in the silicon bulk are given as:
V(x=0)
=
Vs
(3.18)
V(X=o)
=
0
(3.19)
E(x=0)
=
Es
(3.20)
E (x=o)
=
0
(3.21)
P(x=0)
=
ps
(3.22)
-q^s
=
PBexp(
)
(3.23)
kBT
P (X=co)
=
PB
(3.24)
qvF
=
Hi exp(
)
(3.25)
kBT
N(x=0)
=
Ns
(3.26)
q^s
=
NBexp(
)
(3.27)
kBT
N (X)
=
Nb
(3.28)
-qvF qvBE
=
r^exp (
exp ( )
(3.29)
kBT kBT
Using the above boundary conditions and solving Eq.(3.17) by quadrature,
d2V/dx2 = (dV/dx)(d/dV)(dV/dx)
= (1/2)(d/dV)(dV/dx)2
= (1/2) (dE2/dV)2
(3.30)


CHAPTER 8
PHYSICS-BASED TIME-TO-FAILURE EXTRAPOLATION ALGORITHM
USING CURRENT-ACCELERATED CHANNEL HOT CARRIER STRESS
8.1 Introduction
Voltage-acceleration during channel hot carrier (CHC) stress has been employed
industry-wide to determine the maximum operation voltage permitted for a metal-oxide-
silicon transistor (MOST) to meet the operation ten-year time-to-failure (TTF0P)
requirement [33-35]. In this method, the failure rate of TTF data obtained at higher
stress voltages and short stress times are extrapolated to determine the TTF0P at the
desired lower operation voltage. It is based on the assumption that degradation
mechanism at the higher stress voltages applied to MOSTs drain/body junction and
bipolar junction transistors (BJTs) emitter/base junction are independent of the
magnitude of the stress voltages. Transistors degradation has been attributed to the
generation of oxide/silicon interface traps, and charging & generation of oxide traps by
the energetic or hot electrons and holes [111-116], The voltage-extrapolation from
high-stress voltage tacitly assumes that the fundamental failure mechanisms due to hot
carriers are independent of the applied voltages which is untenable because the
degradation mechanisms have hot-carrier kinetic energy or applied voltage threshold
[38, 23, 109], below which the degradation mechanisms cease. To characterize the
degradation mechanisms of BJT technologies at the operation voltage (<3.3V) with
95


85
Figure 7.1 Cross-sectional view of a reverse biased p/n+ junction covered by Si02
showing three dominant bond-breaking pathways to generate interface
traps, (a) Cross-sectional view, (b) Transition energy band diagram along
the oxide/silicon interface. Bond breaking is marked star and vertical
dash line. Open triangles are interface traps. Filled triangles are Si:H and
SiO:H. Dots are electrons and circles are holes. Electron and hole motions
are marked by arrows with solid and open tips, respectively. Adapted from
Sah [42, p. 123].


REFERENCES
[1] Semiconductor Industry Association, "The National Technology Roadmap for
Semiconductors," p.81, 1994.
[2] C.-T. Sah, "Evolution of the MOS Transistor From Conception to VLSI,"
Proceedings of the IEEE, Vol.76, No. 10, pp. 1280-1326, Oct. 1988.
[3] H-N Yu, A. Reisman, C.M. Osburn, and D.L. Critchlow, "ljim MOSFET
VLSI Technology: Part 1 An Overview," IEEE Transaction on Electron
Devices, Vol.26, No.4, pp.318-324, Apr. 1979.
[4] Semiconductor Industry Association, "The National Technology Roadmap for
Semiconductors," p.ll, 1994.
[5] M. Bohr, S.H.Ahmed, L. Brigham, R. Chau, R. Gasser, R. Green, W.
Hargrove, E. Lee, R. Natter, S. Thompson, K. Weldon, and S. Yang, "A High
Performance 0.35|im Logic Technology for 3.3V and 2.5V Operation," IEDM,
No. 10.2.1, pp.273-276, Dec. 1994.
[6] M. Bohr, S.S. Ahmed, S.U. Ahmed, M. Bost, T. Ghani, J. Greason, R.
Hainsey, C. Jan, P. Packan, S. Sivakumar, S. Thompson, J. Tsai, and S. Yang,
"A High Performance 0.25|im Logic Technology Optimized for 1.8V
Operation," IEDM, No.33.1.1, pp.847-850, Dec. 1996.
[7] K.F. Lee, R.H. Yan, D.Y. Jeon, G.M. Chin, Y.O. Kim, D.M. Tennant, B.
Razavi, H.D. Lin, Y.G. Wey, E.H. Westerwick, M.D. Morris, R.W. Johnson,
T.M. Liu, M. Tarsia, M. Cerullo, R.G. Swartz, and A. Ourmazd, "Room
Temperature 0.1 |im CMOS Technology with 11.8ps Gate Delay," IEDM,
No.6.5.1,pp. 131-134, 1993.
[8] B. Davari, R.H. Dennard, and G.G. Shahidi, "CMOS Scaling for High
Performance and Low Powers The Next Ten Years," Proceeding of the IEEE,
Vol.83, No.4, pp.595-606, Apr. 1995.
[9] L. Su, S. Subbana, E. Crabbe, P. Agnello, E. Nowak, R. Schulz, S. Rauch, H.
Ng, T. Newman, A. Ray, M. Hargrove, A. Acovic, J. Snare, S. Crowder, B.
Chen, J. Sun, B. Davari, "A High-Performance 0.08|im CMOS," Symposium
on VLSI Technology Digest of Technical Papers, No.2.1, pp. 12-13, 1996.
151


120

^^eleetran-3-13V
r
KEe|ectron4.25eV

^ ^electron-3 37e V
4-
T)
i Eewp1.12eV
Ofi

Q
KEhole=4.25eV
O)

(2) (3)
'
(4)
Figure 9.1 Four fundamental 4.25 eV hot hole generation pathways initiated by a
primary electron. KEelectron is the threshold kinetic energy of the primary
electron. (1) Hot hole generation by Auger recombination of a hot electron
with a thermal hole. (2) Hot hole generation and thermal electron
generation by a Auger recombination of a hot electron with a thermal hole.
(3) Hot hole generation by interband-impact of a hot electron with a
thermal hole, transferring the energy to the thermal hole. (4) Hot hole
generation and thermal electron generation by interband-impact of a hot
electron with a deep valence band electron.


67
interface traps along the channel and near to the drain region of a nMOST, the
magnitude of IB.1C decreases when the electronic boundary of the surface drain/body
space-charge-layer extends towards the center of the channel by a positive VDX. The
decrease in IB_ 1C is caused by the decreased number of interface traps available for
recombination in the portion of channel modulated by VDX. Figure 5.1(b) shows the
five variations of the IB current with VDX for five possible NIT spatial distributions.
Curve (0) is the IB characteristic for a fresh unstressed nMOST. For low and uniform
Nit distribution, IB decreases with a small constant slope when VDX increases. Curve
(1) shows a non-uniform NIT distribution within the drain/body surface space-charge-
layer, i.e. 0 < y < yD, and uniform NIT distribution along the channel. Curve (2) exhibit
a constant AIB/gVDX slope and hence indicates a uniform N1T distribution along the
channel. For VDX < 0, the constant IB also suggests uniform NIT distribution in 0 < y <
yD. Curve (3) shows that IB decreases with a varying slope for VDX > 0 V and thus
indicates non-uniform NIT distribution. A uniform NIT distribution is implied in 0 < y <
yD by using the same interpretations given above. Curve (4) illustrates the case for a
non-uniform NIT distribution along the channel and also within the drain surface space-
charge-layer.
5.3 Experiment and Results
Industrial n-channel MOST is used in the demonstration of this DCIV Profiling
technique. The n-MOST is fabricated in a p-well with n-epilayer/p-substrate and has
drawn channel length L=1.6|im, drawn channel width W=100.0fim, gate oxide
thickness xOx=160A, and p-well doping concentration NAA=1.4xl016cm~3. The SE-


43
Vs = 2Vf-Vbe+
kBT
log.
(^GB ^FB Vs) 2
(4kBT/q)VAAAI
(3.35)
In Eqns. (3.33)-(3.35), the iterative correction terms (Aa, Ad and Aj) are
Aa = 1-US-[exp (Us)-1-US] exp (-2Uf+Ube) (3.36)
Ad = 1-exp (-Us) [exp (us)-1-US] exp (-2Uf+Ube) (3.37)
Aj = 1- (1+US) exp (-Us) + [exp (-2US) -1+US] (3.38)
Using Eqns.(3.31)-(3.33), the relationship between \|/s and VGB is computed for
x0X=15nm, NAA=1016cm-3, VFB=0V, and T=300K. The result is shown in Figure 3.8
with VBE as parameter. In the accumulation and depletion ranges, the results computed
for non-zero VBEs are similar to the case with VBE=0 V. This indicates that, under
low-to-moderate forward-biases, the excess surface electron concentration due to
electron injection from the forward-biased emitter junction is not appreciable to affect
the field and potential distribution in the surface space-charge-layer. However, in the
strong inversion range, excess surface electron concentration reduces the amount of
surface band bending by VBE.
Vs, inv 2Vf Vbe (3.39)
The results shown in Figures 3.6 and 3.7 are combined and replotted in Figure 3.8
which illustrates the gate-voltage dependence of surface recombination rate, Rss, for
four forward-bias conditions (VBE=0.2, 0.3, 0.4 and 0.5V). This theoretical calculations
indicates that base current, measured during DCIV, will exhibit a maximum when its
surface recombination components dominate. For a single energy-level interface trap,


45
VGB/(1V)
Figure 3.8 Relation between surface recombination rate (Rss) and applied gate voltage
(VGB) under equilibrium (VBE= 0 V) and non-equilibrium (VBE= 0.2, 0.3,
0.4 and 0.5 V) conditions. Parameters used: xox=15 nm, NAA=1016cm-3,
VFB=0 V, Nn-1010 cm'2 and T=300K.


D-sar 1 D-satO
114
102
101
10
10"1
10
10-1
1CT2
1CT3
10 101 102 103
%A\
B-DE-peak
Figure 8.8 Dependence of -AID_sat/ID_sat0, AVGX on AIB.peak/IB.peak for 0.35 pm
nMOST under CHE (open and filled triangles) and CACHE (open and
filled circles). CHE stress condition: VDS= 3.5 V, VGS= 0.5V, VSB= 0.0V.
CACHE stress 'condition: VDS= 3.5V, VGS= 0.5V, VSB= -0.8V.
AVgt/(1V)


79
Vd=Vg/(1 V)
Figure 6.2 Forward and reverse IDsat characteristic of a nMOST stressed by channel
hot electrons. Difference between the post-stress forward (F) and reverse
(R) characteristics reveal the asymmetrical nature of channel hot electron
stress. tstress= 20 ksec. Pre-stress forward and reverse IDsat characteristics
are identical.


96
relatively short turn-over time, Neugroschel and Sah proposed two methods to increase
the stress current [39-40] at low stress voltage, as contrast to high voltage stress, to
accelerate the degradation rate of the transistors, and using the DCIV measurement to
monitor the hot carrier-induced degradation. In this chapter, we will demonstrate that
the BJT current-acceleration methods can be extended to MOSTs under CHC stress.
First, long channel MOSTs structure will be used for demonstrating the current-
accelerated channel hot carrier stress methodology. Long channel MOSTs are used
because the transistor is essentially 1-D and this will simplify and aid in the analysis.
Upon laying out the framework and understanding the basis physic of the current-
accelerated channel hot carrier stress (CACHC) methodology in long channel MOSTs,
short channel MOSTs will be employed to demonstrate the operation TTF
determination for state-of-the-art MOSTs. A physics-based TTF extrapolation
algorithm is presented. A comparison with the traditional voltage-accelerated TTF
extrapolation using the short channel nMOST results will also be presented.
8.2 Methodology of Current-Acceleration Channel-Hot-Carrier Stress
As delineated by Sah and Lu [23, 109], the kinetic of fundamental degradation
mechanisms in MOSTs and BJTs depends on (1) density and (2) kinetic energy of hot
carriers, (3) the threshold kinetic energy of degradation pathway and the density of
residual-fabrication created (4) active or non-hydrogenated and (5) passive or
hydrogenated (Si:H and Si:OH) oxide and interfacial electron and hole traps. Transistor
degradation rate may be increased by increasing the hot carrier kinetic energy using
higher stress voltage (the traditional voltage acceleration method) or larger hot carrier


146
11 rirrm 11111111111111 n 11111111 mi i rmmi 11 m 111 n 11111111111111111 n
p+Substrate
J Substrate
Figure 10.4 Cross-sectional view of a pMOST under CHH stress. The hydrogen
diffusion pathways from the broken interfacial bonds and dehydrogenated
boron are shown. The dotted line indicates the edge of the space-charge
layer.


157
[67] M.Y. Ghannam, R.P. Mertens, R.F. De Keersmaecker, and R.J. Van
Overstraten, "Electrical Characterization of the Boron-Doped Si-Si02
Interface," IEEE Transaction on Electron Devices, Vol.32, No.7, pp.1264-
1271, Jul. 1985.
[68] W. Shockley and W.T. Read, "Statistics of Recombinations of Holes and
Electrons," Physical Review, Vol.87, pp.835-842, Sep. 1952; R.N. Hall,
"Germanium Rectifier Characteristics," Physical Review. Vol.83, pp.225, Jul.
1951.
[69] Chih-Tang Sah, Florida Solid-State Electronic Laboratory Seminar, Dec 20
1996. Also see Chih-Tang Sah, Fundamentals of Solid-State Electronics -
Solution Manual, section (910) on p.103, World Scientific Publishing Co.
1996. New Jersey and Singapore.
[70] Chih-Tang Sah, Fundamentals of Solid-State Electronics Study Guide,
Section 411, p.129, World Scientific Publishing Co. Dec 1993. New Jersey
and Singapore.
[71] C.-T. Sah, K.M. Han, and D.O. Martin, "DC Current-Voltage Diskette for
Submicron Silicon Transistor Reliability Characterization," VI.2, Oct 1996.
[72] C.-T. Sah and T. Nishida, "Mechanisms of Electronic Trapping in Si02 on Si,"
The Physics of Semiconductor. Singapore; World Scientific, vol.l, pp.28-40;
also presented at the Plenary Presentation, 21th Int. Conf. Physics of
Semiconductors, Beijing, Aug. 10-14, 1992.
[73] A.S Grove and D.J. Fitzgerald, "Surface Effects on P-N Junctions:
Characteristics of Surface Space-Charge Regions Under Non-Equilibrium
Conditions," Solid-State Electronics, vol.9, pp.783-806, 1966.
[74] Chih-Tang Sah, "Fundamental of solid-state electronics," World Scientific
Publishing Co,. Singapore 1991. See Eqn.(683.10) on p.666.
[75] D.R Young, E.A. Irene, D.J. DiMaria, and R.F. Dekeersmaecker, "Electron
Trapping in Si02 at 295 and 77K," J. Appl. Phys., vol.50, no. 10, pp.6366-
6372, Oct 1979.
[76] H. Haddara and S. Cristoloveanu, "Two-Dimensional Modeling of Locally
Damaged Short-Channel MOSFETs Operating in the Linear Region," IEEE
Trans. Electron Devices, vol.34, no.2, pp.378-385, Feb. 1987.


53
Electrons
Figure 4.3 Transition energy band diagram of BiMOS substrate hot electron injection.


D-sat' 1 D-satO
81
10
1CT1
icr2
10"3
10~4
%AI
B-DE-peak
Figure 6.4 Dependence of -AID_sat/ID.sat0, AVGT and AS on AIB.peak/IB_peak0 in 0.35^im
silicon nMOSTs stressed by channel hot electrons.
A,0,0: Vds/Vgs=4.0/1.9, 3.75/1.8, 3.3/1.6.
A,Vds=4.0V, Vgs= 1.0, 2.0, 3.0, 4.0V
(A/SSV 0AV


156
[56] T.Y. Chan, P.K. Po and C. Hu, "Dependence of Channel Electric Field on
Device Scaling." IEEE Electron Dev. Lett., Vol.6, No. 10, pp.551, Feb 1985.
[57] Y. Nissan-Cohen, "A Novel Floating-Gate Method for Measurement of Ultra-
Low Hole and Electron Gate Currents in MOS Transistors," IEEE Electron
Dev. Lett., Vol.7, No.10, pp.561, Mar 1986.
[58] N.S. Saks, P.L. Heremans, L.V.D. Hove, H.E. Maes, R.F.D. Keersmaecker and
G.J. Declerck, "Observation of Hot-Hole Injection in NMOS Transistors Using
a Modified Floating-Gate Technique," IEEE Trans. Electron Devices, vol.33,
no.10, pp. 1529, May 1986.
[59] Chih-Tang Sah, "Effect of Surface Recombination and Channel on P-N
Junction and Transistor Characteristics," IRE Transaction on Electron Devices,
Vol.9, No.l, pp.94-108, Jan. 1962.
[60] Chih-Tang Sah, "A New Semiconductor Tetrode The Surface-Potential
Controlled Transistor," Proceeding of the IRE, Vol.49, No.11, pp.1623-1634,
Nov. 1961.
[61] J.T. Kavalieros and C.-T. Sah, "Separation of Interface and Nonuniform Oxide
Traps by the DC Current-Voltage Method," IEEE Transaction on Electron
Devices, Vol.43, No.l, pp. 137-141, Jan. 1996.
[62] A. Neugroschel, C.-T. Sah, and M. Carroll, "Degradation of Bipolar Transistor
Current Gain By Hot Holes During Reverse Emitter-Base Bias Stress," IEEE
Transaction on Electron Devices, Vol.43, No.8, pp. 1286-1290, Aug. 1996.
[63] C.-T. Sah, A. Neugroschel, K.M. Han, and D.O. Martin, "The DC Current-
Voltage Method for Measuring Interface Traps in Submicron Silicon
Transistors A Demonstration," TECHCON96, Phoenix, Arizona, USA.
[64] Arnost Neugroschel and Chih-Tang Sah, "Fundamental Physical Degradation
Mechanisms in Silicon Bipolar Junction Transistors," Publication C94321,
Semiconductor Research Corporation, Research Triangle Park, NC.
[65] C.-T. Sah, R.N. Noyce, and W. Shockley, "Carrier Generation and
Recombination in P-N Junctions and P-N Junction Characteristics,"
Proceeding of the IRE, Vol.45, pp. 1228-1243, Sep. 1957.
[66] M. W. Hillen and J. Holsbrink, "The Base Current Recombining at the
oxidized silicon surface," Solid-State Electronics, vol.26, pp.453-463, May
1983.


CHAPTER 11
SUMMARIES AND CONCLUSIONS
Channel-hot-carrier induced degradation in silicon-based MOSTs is a major
constraint for the long-term reliability of VLSI/ULSI circuit. As the MOST continues
to reduce to smaller physical dimensions at constant or near constant supply voltages,
the magnitude of the channel electric field increases. The increased electric field
accelerates the electrons in nMOST and holes in pMOST to high kinetic energy which
can generate and charge oxide traps in the gate oxide and interface traps at the Si02/Si
interface. These electrically-active oxide and interface traps degrade the performance of
the MOSTs after long-term operation.
In this thesis, the electrical degradation characteristics of MOSTs under channel
hot carrier stress was investigated using a novel Direct-Current Current-Voltage (DCIV)
characterization technique. This technique employs the surface recombination current
of the parasitic gate-controlled BJT in MOST structure as a degradation monitor. Two
applications are illustrated: (1) separation of stress-generated oxide and interface traps
in submicron MOSTs and (2) spatial profiling of the stress-generated interface trap
distribution. The stress-generated surface recombination current is shown to exhibit a
linear correlation with the traditionally measured transistors parameters (IDsat, VGT, and
S). A simple drain-current degradation model was successfully used to account for the
observed correlation. This correlation provides a basis for using the DCIV method to
148


23
Figure 2.9 shows the second current-accelerated which forward-biases the
n+substrate/p-body junction. In this case, a negative voltage is applied to the n+
substrate of the n-channel MOST and the p-body and n+source are grounded during
CACHE stress.


101
Figure 8.3 (a) Time-to-Failure versus drain stress voltage of 100(imx 100|im nMOST
for no-current-acceleration stress (four circles) and current-accelerated
stress (five solid dots at VBE= 0.63 V, one solid triangle at VBE= 0.74 V
and an asterisks at VSB= -0.7 V. All transistors are stressed at VCS=0.9V.
The inset shows the DE-DCIV (Ib'Vgb) characteristics (measured at
Vdb=-0-4V, Vsb=Veb=0.0V) as a function of CHE stress time (0, 85.3,
738.7, 2319.7, 17227.6, and 63758.8 sec). Stressed at VDS= 7.0 V, VGS=
0.9 V, and VSB=VBE= 0 V.


90
gap potential and VFN+ is the Fermi level potential above the conduction band edge of
the quasi-neutral n++ drain-emitter. The corresponding normalized interface trap
generation efficiency is then given by:
Wito = exP[-,,l (SE/qN^) / X2] [jVR+VBI-Vsi_Gap-VFN+ -
VR+VBi-Vsi_Gap-VFN+_ VBB J (7-14)
As shown in Eqns. (7.12) and (7.14), the interface trap generation efficiency has a
strong voltage dependence on the applied stress voltage when the electron (or hole)
enters the space-charge-layer from the high-field edge. On the contrary, if the electron
(or hole) enters from the low-field edge of the space-charge-layer, the interface trap
generation efficiency has no voltage dependence, as seen from Eqn.(7.10). Hence, a
voltage-dependency experiment can be designed to correlate the experimental results
with the interface trap generation theory presented in this section.
7.3 Experiment and Results
Four sets of Time-to-Failure SAM experiments on silicon nMOST, pMOST and
npnBJT were analyzed using the new interface trap generation theory. The nMOST
data are obtained from a channel-hot-electron (CHE) stress experiment where the
nMOST test structure was fabricated from a 1-micron BiCMOS technology with 100
|im channel length, 100 p.m channel width, and 160 A thick gate oxide. The pMOST
data are similarly obtained from a channel-hot-hole (CHH) stress experiment where the
pMOST test structure was fabricated from a 0.8-micron CMOS technology with 20 |im
channel length, 20 |im channel width, and 110 A thick gate oxide. The experimental


-12
124
Figure 9.2 Drain-emitter DCIV (DE-DCIV) characteristics of L/W=100 (im/100 fim
nMOSTs, measured at ^ebd- 0-4 V and VBS= 0 V after CHE stressed at
VDS= 7.0 V, VGS= 0.9 V, and VBS= 0 V (B=Body=X) for 0, 0.1 k, 1 k, 11 k
and 26 k seconds. The inset shows the exp(qVBD/nkBT) dependence of the
stress-induced body current, AIB= Ib(Vgb= 2.5 V, tstress= 2.6 ksec) -
Ib(VGb= 2.5 V, tstress= 0), indicating n=2 due to recombination in the drain-
junction space-charge region.


68
DCIV configuration is used to inject minority carriers (electrons) into the p-body from
the forward-biased source junction. In this illustration, we generate non-uniform NIT by
stressing the nMOST under CHE stress condition (VDS=8.0 V, VGS=4.0 V, and
Vsx=0.0 V) for 120 minutes.
Figure 5.2 shows the post- and pre-stressed DE- and SE-DCIV characteristics.
Both DE- and SE-DCIV are measured at VBE=0.4V and VCB=0.0V. For the case of SE-
DCIV measurement, the post-stressed IB increase for IV < VGB < 0V is attributed to
stress-generated interface traps along the channel which results in more electron-hole
recombination through these additional interface traps in the channel. New bulk defects
are not created by the CHE stress; otherwise, the measured post-stressed IB will show a
relative constant increase over the un-stressed IB-VGB data because the bulk diffusion-
recombination IB components (IB.b2 and IB_bl in Figure 3.2) are independent of surface
modulation by the applied VGB. There is also negligible shift of the post-stressed IB
peak, in comparison with the unstressed IB peak at VGB=-0.35V, along the voltage-axis.
This suggests that there is negligible built-up of oxide charge during the 120-minutes
CHE stress. Since the channel hot electrons experience the largest accelerating electric
field near the drain/body metallurgical junction, we anticipate that there should also be a
spatial distribution of stress-generated interface within the surface drain space-charge-
layer. Thus the drain surface-charge-layer recombination current IB_2S, as discussed in
chapter 3, should dominate when the drain is used as an emitter in the DCIV
measurement. Indeed, this was observed as indicated by post-stressed DE-IB-VGB in
Figure 5.2.


121
K^electron^ ^X-h + EG EG (9-2)
= 4.12 eV.
Pathway (3) shows the hot hole generation by interband impact of a hot electron
with a thermal hole where the hot electron transfers all its kinetic energy to the thermal
hole. For this pathway, the threshold kinetic energy KEelectron.3 is simply the hole
barrier height which is given by
KEelectron-S = ^X-h (9-3)
= 4.12 eV.
Pathway (4) shows the hot hole generation by a hot electron impacting a deep
valence band electron with a thermal electron to raise this valence electron to the
conduction band edge. For this interband-impact pathway, the threshold kinetic energy
KEelectron^ is given by
KEelectron^ = tX-h + EG (9-4)
= (4.12+1.12) eV
= 5.37 eV.
Thus, among these four pathways, there are three distinct threshold electron kinetic
energies: 3.13 eV, 4.12 eV and 5.37 eV to create a 4.25 eV hot hole [126]. These
threshold electron kinetic energies are derived based on energy conservation without
considering momentum conservation. Lu and Sah have shown that energetic carrier
generations in silicon via the impact-Auger transitions can be analyzed using energy
conservation alone [109]. By including the silicon band structure and Bragg reflection
or umklapp process in their calculations, they showed that the threshold electron


33
pairs prevail in these regions but can be neglected because the large base recombination
current will overwhelm this small generation current even at small forward bias [96].
Thus, the measured IB in the BE configuration consists of only two emitter components
and three base recombination components. The two emitter components are the e-h
recombinations at the recombination centers in the n+-substrates bulk-quasi-neutral
region (iB-BE-el-bulk) ar,d at the interface between the n+ substrate and metal contacts
(^B-BE-ei-contact) respectively. The three base e-h recombination components are: (1)
IB-be-c-1 over the n-channel-interface which is same as Ib-de-c-1 (2) lB-BE-b2 in the
n+substrate/body junction bulk space-charge region, (3) Ib-be-c-t same as (1) but
outside of the n-channel-perimeter. Thus, the measured IB.BE IS given by
Ib-BE = ^B-BE-el-bulk + ^B-BE-el-contact + ^B-BE-cl
+ ^-BE-bl + ^B-BE-cl (3-2)
In (3.2), only Ib-be-c1 is modulated by the gate voltage. Thus the BE-DCIV
measurement configuration gives greatly simplified data analysis and provides the
unique locationing of the N[T (n=l) in the channel region. This BE configuration had
been used by other researchers to study the surface recombination rate or velocity S0 at
the Si02/Si interface [66, 67] and the areal uniformity of positive oxide charge build-up
and interface trap generation using substrate hot-electron injection [61].
3.4 Theory of Surface Recombination Rate
The surface recombination current components discussed in section 3.3 can be
evaluated by extending the Shockley-Read-Hall (SRH) recombination kinetic for a
single trap level [68] to determine the surface recombination rate Rss of electron and


84
based on a physical bond-breaking mechanism and considers the bond-breaking
efficiency of the hot electrons and holes along the surface channel on oxidized silicon.
It identifies the primary and dominant hot carriers (electrons or holes) that are involved
in the interface trap generation process. In this chapter, the analytical interface trap
generation theory is presented and the validity of the model is supplemented by
experimental verification.
7.2 Theory
Consider a n-channel MOS transistor biased in the flat-band condition shown in
Figure 7.1(a), the drain n+/p junction is reverse biased by VR which results in a wider
space-charge-layer (SCL). The corresponding energy band diagram is depicted in
Figure 7.1(b). The drain junction space-charge-layer can accelerate electrons, injected
at the low-field edge (LFE) of the drain SCL, and holes, injected at the high-field edge
(HFE) of the drain SCL, to higher kinetic energy when the electrons or holes transit
across the SCL. When the accelerating electrons and holes acquire sufficient energy,
greater than the bond-breaking energy (qVBB), they can break the hydrogenated
interfacial bonds (Si:H and SiO:H) to release the hydrogen and create new interface
traps.
Based on Sahs interface trap generation model, the rate of interface trap
generation (dN1T/dt) is proportional to the injected carrier flux density (Jj/q), the
probability of surviving a scattering, and the probability of bond-breaking. Thus, the
interface trap generation rate is given by the product of these three terms:


88
is zero if the applied voltage is below. The normalized interface trap generation
efficiency for this LFEi pathway is.
Hito = exp [-j (2£/qNAA) /2\\ VBB (7.10)
Next, consider the hole injection pathway from the high-field drain edge labelled
HFEi in Figure 7.1(a). The sources of these holes can be due to thermal generation or
from the interband-impact generation process initiated by the energetic electron when
the electron kinetic energy, KEelectron, is greater than the electron-hole pair generation
threshold (EPN > EG.Si 1.17eV) [109]. Typically, the concentration of the Electron-
Impact-Generated (EIH) holes is higher than the thermally-generated hole
concentration. These impact-generated holes can be back-injected into the drain space-
charge layer from the maximum-field edge and break the interfacial bonds when its
energy exceeds the bond-breaking energy. The bond-breaking distance for this HFHi
interface trap generation pathway is:
BB
-
(2£/qNAA) [i VR+VBI i
Vr+VBi Vbb ]
(7.11)
Similarly, the corresponding normalized interface trap generation efficiency is given by.
H ito exp [ t| (2 £ / ) /z^] C VR+VBi 1 VR+VBI VBB ] (7.12)
where VBI is the built-in potential of the abrupt n+/p drain junction. Comparing with
Eqns.(7.9) and (7.10), eqns.(7.11) and (7.12) reveal that this HFHi interface trap
generation pathway has a very strong dependence on the applied stress voltage, VR.
This characteristic strong voltage dependence of dBB or rj1T0 provides a feature for the
theory to be experimentally verified. This will be described and elaborate in the


110
When applied to the BJT under emitter-base reverse-bias stress in which VBI is no
longer zero, the threshold drops to zero as shown by Sah [117].


60
Vgb / (1V)
Figure 4.5 Effect of stress on the collector current measured at VEB (bottom-emitter)=
-0.3 V, VCB= 0 V, and 297K. (a) 5xl018 cm-2 SHEi stress at VGB = 7.5 V,
VCB = 4.0V and IG = InA. (b) 1.4xl016 cm-2 SHEi stress at VGB = 12 V,
VCB = 10 V, and IG = InA. (c) CHEi stress at VGB=VDB= 16 V and
floating VSB and ID = 1 (lA for 1 sec (curve 1) and 500 s (curve 2).


49
test structure for accurate measurement of the small-signal capacitance and thus is not
suitable for test structure with very small transistors. For the d.c. ID-VG method, the
extracted threshold voltage shift can be used to monitor the density of the oxide traps,
provided that the oxide trap density is areally uniform and the interface trap density is
sufficiently low. The interface trap density can be qualitatively obtained from the
subthreshold slope change. However, if both oxide and interface traps are generated
during the hot carrier stress experiment, then it is not easy to separate the relative
contribution of QOT and QIT from a single ID-VG measurement because the total gate
voltage shift is due to the build-up of both QOT and QIT.
AVg = AVg_ot + AVg_iT (4.1)
(AQrprp + AQit)
where AVG.OT is a component of the total gate voltage shift, AVG, caused by the
charged oxide trap. By the same argument, the interface trap contributes a AVG.IT
component. In addition, ID also depends on mobility which is changed by QIT and QOT.
In this chapter, we will show that the DCIV method can be used to separate the
AQot and AQIT [41]. This novel method contains two features: (1) the increase in the
magnitude of base current is used to monitor the stress-generated QIT, (2) the lateral
voltage shift of the collector current versus gate voltage is used to detect the QOT.
4.2 Experiment
To demonstrate the separation of QIT and Q0T using the DCIV method, a nMOST,
fabricated with a junction-isolated p-well, is subjected to Substrate Hot Electron


162
[121] Eiji Takeda, Cary Y. Yang, and Akemi Miura-Hamada, "Hot-Carrier Effects in
MOS Devices," pp.75, Academic Press 1995.
[122] Wenliang Chen and Tso-Ping Ma, "Oxide Charge Buildup and Spread-out
During Channel-Hot-Carrier Injection in NMOSFETs," IEEE Electron Device
Letters, Vol.13, No.6, pp.319-4321, Jun 1992.
[123] Werner Weber, Martin Brox, Roland Thews, and Nelson Saks, "Hot-Hole
Induced Negative Oxide Charge in n-MOSFETs," IEEE Trans. Electron
Devices, Vol.42, No.8, pp. 1473-1480, Aug 1995.
[124] See Figure B2.5, on p.411 and Table B2.2 on p.409 of Chih-Tang Sah,
Fundamental of Solid-State Electronic Study Guide, World Scientific
(Singapore), 1993.
[125] T.H. Ning, "Capture cross section and trap concentration of holes in silicon
dioxide," Journal of Applied Physic, Vol.47, No.3, pp.1079-1081, Mar 1976.
[126] See Section 360, pp 270 of Chih-Tang Sah, Fundamental of Solid-State
Electronic, World Scientific (Singapore), 1991.
[127] Eiji Takeda, Yoshinobu Nakagome, Hitoshi Kume, Norio Suzuki, and Shojiro
Asai, "Comparison of Characteristics of n-Channel and p-Channel MOSFETs
for VLSIs," IEEE Trans. Electron Devices, Vol.30, No.6, pp.675-680, Jun.
1983.
[128] Andreas Schwrin and Wilfred Hansch and Werner Weber, "The Relationship
Between Oxide Charge and Device Degradation: A Comparative Study of n-
and p-Channel MOSFETs," IEEE Trans. Electron Devices, Vol.4, No. 12,
pp.2493-2500, Dec. 1987.
[129] Tong-Chern Ong, Ping-Keung Ko and Chenming Hu, "Hot-Carrier Current
Modeling and Device Degradation in Surface-Channel p-MOSFETs," IEEE
Trans. Electron Devices, Vol.37, No.7, pp. 1658-1666, Jul. 1990.
[130] F. Matsuoka, H. Iwai, H. Hayashida, K. Hama, T. Toyoshima, and Kenji
Maeguchi, "Analysis of Hot-Carrier-Induced Degradation Mode on
pMOSFETs," IEEE Trans. Electron Devices, Vol.37, No.6, pp.1487-1495,
Jun. 1990.
[131] Toshiaki Tsuchiya, Yukio Okazaki, Masaysu Miyake and Toshio Kobayshi,
"New Hot-Carrier Degradation Mode and Lifetime Prediction Method in
Quarter-Micrometer PMOSFET," IEEE Trans. Electron Devices, Vol.35,
No. 12, pp.404-408, Feb. 1992.


150
hot-hole injection pathways are delineated. Two pathways are experimentally observed:
hot-electron/hot-hole interband impact pathways in high-voltage long channel nMOST
with electron kinetic energy threshold of 5.12 eV, and hot-electron/thermal-hole Auger
recombination pathway in low voltage short-channel nMOSTs with electron kinetic
energy threshold of 3.13 eV. The positive oxide charge buildup rate is anticipated to
decrease drastically when the supply voltage is decreased to less than 3.3V. Thus
interface trap generation is expected to be a dominating degradation mechanism in
future generation of deep-submicron MOST.
For p-channel MOS transistor during channel hot hole stress, interface traps are
generated and occurs mainly near the drain junction. A reduction of the interface trap
density over the p-channel of MOS transistors during stress is observed for the first
time. Hydrogen released from the hydrogenated boron in the p+drain region by the
channel hot holes is proposed. The released hydrogen diffuses from the p+drain into the
channel region and passivates fabrication-residual interface traps over the channel and
source interface.


147
(10.3)
(10.2)
= 500
Lch is the channel length and x0X is the oxide thickness. Doxjde/Dinterface= 1/2 is
assumed to account for the slightly larger hydrogen diffusivity along the Si02/Si
interface.
10.4 Summary
In summary, we have shown that interface traps are generated in p-channel MOS
transistor during CHH stress, mainly near the drain junction. During the CHH stress,
mobile hydrogen is released from hydrogenated boron in the p+drain by the hot holes
injected from the p-channel. The released hydrogen diffuses from the p+drain into the
channel region and passivates residual fabrication interface traps over the channel and
source junction.


134
5.37 eV interband-impact generation threshold discussed in Section 9.2. Thus, hot hole
injection into Si02 via the interband-impact pathway dominates when the electron
kinetic energy exceeds the 5.37 eV threshold i.e. q(VD VDsat) > 5.37 eV.
Assuming VDsat = VGS VGT, where VGT is the threshold voltage of the nMOST,
then, a high-VD and low-VG CHE-stress condition would satisfy the (VD VDsat) = (VD
- VG + VGT) > 5.37 V to positively charge the Q0T near the drain space-charge-region
by the CHE-impact-generated hot holes. This could account for the high-VD and low-
VG stress conditions used in [55, 57-58, 113, 119-123] to inject holes into the Si02
during CHE stress.
9.4.3 +0OT Build-up in Short-Channel nMOSTs
From the long-channel MOSTs results just presented, interband-impact generation
of secondary hot holes by primary channel hot electrons was shown to be a primary
cause of +QGt build-up in nMOSTs when VD VDsat > 5.37 V. However, as discussed
in section 9.2, there are three other pathways with lower electron threshold kinetic
energies, among the four impact-Auger transitions, to give the 4.25 eV barrier
surmounting hot holes. These are the hot-electron/thermal-hole impact collision
pathway and the hot-electron/thermal hole Auger recombination pathway, which give an
electron threshold of 4.25 eV, and the hot-electron/thermal-hole Auger recombination
pathway which gives an electron electron threshold of 3.13 eV. These lower thresholds
for +Qot build-up is of great practical importance. It indicates that +QGt would cease
to be an important limitation of current and future shorter channel and thinner gate
oxide transistors designed for operation at lower voltages. This is illustrated by the


87
'IT
r dNTT -|
rJi i
/

L dt J
L q J
= exp
^BB
*3
2
^BB
Hito
(7.6)
(7.7)
(7.8)
^BB
where riIT0 = exp(-dBB/X2) is the normalized interface trap generation efficiency.
Analytical expressions for voltage dependence of the bond-breaking distance, dBB, can
be derived by using the depletion approximation of an abrupt n+/p junction and by
tracing the three bond-breaking pathways, illustrated in Figure 7.1, of the energetic
electrons and holes.
First, let consider the electron injection from the low-field edge of the drain space-
charge layer (the LFEi pathway shown in Figure 7.1), the accelerating electron in the
drain space-charge layer can break an interfacial bond when it has kinetic energy greater
than the bond-breaking energy, qVBB. Thus, dBB can be expressed as follows:
'tBB
= 1 (28/qN^) ( VBB ) (7.9)
where 8 is the silicon dielectric constant, q is the electronic charge, and NAA is the
acceptor doping concentration. For this LFEi interface trap generation pathway, the
bond-breaking distance is independent of the applied voltage and only depends on the
acceptor doping concentration, NAA, and the threshold bond-breaking energy, qVBB.
This implies that the efficiency of interface trap generation by these energetic hot
electrons is a constant when the applied voltage exceeds the bond-breaking energy, and


BIOGRAPHICAL SKETCH
K. Michael Han was born in 1966 in Singapore. He received a B.S.E.E (high
honors) and a M.S.E.E. degree from the University of Florida in 1990 and in 1992,
respectively. In 1993, he worked at TECH Semiconductor (Singapore), a joint venture
of Texas Instrument, Singapore Economic Development Board, Cannon and Hewlett-
Packard, as a Quality Assurance Engineer. He then returned to University of Florida in
1994 to continue his education under the supervision of Prof. Chih-Tang Sah at the
Florida Solid-State Electronics Laboratory. He will complete his Ph.D. degree in
electrical engineering at the University of Florida in December 1997.
164


159
[87] Wenliang Chen and Tso-Ping Ma, "A New Technique for Measuring Lateral
Distribution of Oxide Charge and Interface Traps Near MOSFET Junctions,"
IEEE Electron Device Letters, vol.12, no.7, pp.393-395. Jul. 1991.
[88] Wenliang Chen, Arthur Balasinski and Tso-Ping Ma, "Lateral Profiling of
Oxide Charge and Interface Traps Near NOSFET Junction," IEEE Trans.
Electron Devices, vol.40, no.l, pp.187-195, Jan. 1993.
[89] J. Stephen Brugler and Paul G.A. Jespers, "Charge Pumping in MOS
Devices," IEEE Trans. Electron Devices, vol.16, no.3, pp.297-302, Mar. 1969.
[90] Alexander B.M. Elliot, "The Use of Charge Pumping Currents to measure
Surface State Densities in MOS Transistors," Solid-State Electron., vol.19,
no.3, pp.241-247, Mar. 1976.
[91] U. Cilingiroglu, "A General Model for Interface-trap Charge-Pumping effects
in MOS Devices," Solid-State Electron., vol. 28, no. 11, PP- 1127-1142, Nov.
1985.
[92] Guido Groeseneken, Herman E. Maes, Nicolas Beltran, and Roger F. De
Keersmaecker, "A Reliable Approach to Charge-Pumping Measurements in
MOS transistors," IEEE Trans. Electron Devices, vol.31, no.l, Jan. 1984.
[93] G. Rupprecht, "Measurement of Germanium Surface States by Pulsed Channel
Effect," Phys. Rev., vol.l 11, No.l, pp.75-81, July 1958; "Cross Sections of
Midgap Surface States in Silicon by Pulsed Field Effect Experiment," J. Phys.
Chem, Solids, Vol. 14, pp.208-213, 1960.
[94] S. Sato, S. Kawaji, and A. Kobayashi, "Deep Trap Levels in Silicon p-n
Junctions," J. Appl. Phys., vol.3, pp.496-497, 1964.; "Surface States on Cleavd
Silicon Surfaces," Surface Science, vol.3, pp.98-100, 1964; "Electrical
Properties of Carrier Generation-Recombination Centers in Silicon p-n
Junctions," J. Appl. Phys., vol.36, no.12, pp.3779-3783, Dec. 1965; and
"Electrical Properties of Surface States on Silicon Surfaces," Surface Science,
vol.4, pp.299-312, 1966.
[95] S. Sato and C.T. Sah, "Measurements of the Thermal-Emission Rates of
Electrons and Holes at the Gold Centers in Silicon Using the Small-Signal
Pulsed Field Effect," J. Appl. Phys., Vol.41, no. 10, pp.4175-4181, Sep. 1970.
[96] C.-T. Sah, A. Neugroschel, K. Michael Han, and Jack T. Kavalieros, "Profiling
interface traps in MOS transistors by the DC current-voltage method," IEEE
Electron Device Letters, vol. 17, no.2, pp.72-74, Feb. 1996.


86
dNIT Jj
^BB
Ysc
exp
^BB
- (y- dy
dt q
^2
23
^BB
(7.1)
The product of the first and second term in front of the integrand of Eqn.(7.1) gives the
decrease of the ballistic hot carrier density due to scattering prior to breaking a bond.
The first term inside the integrand of Eqn.(7.1), exp[-(y-dBB)/A.3], is the fraction of the
injected carrier which has escaped scattering and not broken bond when the carrier
transits from y=(y-dBB) to y=y. (dy/XBB) is the fraction of bond-breaking events within
the dy region. The combined mean-free-path, A,3, contains three components: the
surface scattering, A.s, which includes the phonon, impurity, electron-electron and
electron-hole scattering events; the interband impact-generation of electron-hole pair,
7.pn; and the bond-breaking by the energetic carrier, ?^BB.
1/A3 = 1/As + 1/2PN + 1/XBB (7.2)
B 1/2 + 1/ Abb (7.3)
Since the bond-breaking event occurs infrequently, it is anticipated that A,BB A,PN >
Xs > ^2 > ^3- Solving Eqn.(7.1) and using the approximation (Ysc-dBB)X3,
dN
IT
dt
q
Jt
exp
exp
-d
BB
BB
BB
BB
1-exp
1-
( ^SC ^BB)
(7.4)
(7.5)
The interface trap generation efficiency, r)IT, is defined as the rate of interface trap
generation per injected carrier density. Hence, Eqn.(7.5) can be expressed as follows.


93
npnBJT. For the ITHi npnBJT data, Eqn.(7.14) was used instead to account for the
valence-electron tunneling process. The LSF results gives a bond-breaking energy of
1.877 eV and a standard deviation of 1.41 %. Also included in Figure 7.2 is the
theoretical interface trap generation efficiency for the LFEi pathway, with qVBB=3.1eV
assumed in the calculation.
Albeit the difference in nMOST and pMOST fabrication technology, the LSFs
bond-breaking energies obtained from the nMOST and pMOST data are comparable to
the theoretical 3.101 eV Si:H bond energy [110]. Two implications can be made from
this observation: (1) the origin of the stress-generated interface trap is the Si:H bond at
the Si02/Si interface, and (2) the mechanism of interface trap generation is due to the
energetic holes breaking the weak Si:H bonds during nMOST CHE stress and n/p/n BJT
reverse emitter stress, and the energetic electron breaking the Si:H bonds during
pMOST CHH stress. This bond-breaking interface trap generation process can be
described by the following kinetic equation.
Si: H + c* Si. + Ht + c (7.15)
where c* represents a energetic hot carrier (hole in nMOST and n/p/n BJT and electron
in pMOST), Ht indicates the diffusion of hydrogen away from the broken Si. bond.
The carrier, c, has thermal energy after the bond-breaking process.
For the npnBJTs, the two HFHi and ITHi pathways yield similar bond breaking
energy (qVBB= 1.88 eV) which is expected since the primary carrier responsible in
npnBJT degradation is the energetic hole. The primary difference in bond breaking
energy between the MOST (3.06 eV) and the npnBJT (1.88 eV) is attributed to the


73
VCB/(1V)
CHANNEL y / (urn)
l I I I
0.2 0.3 0.4 0.5
(C)
Figure 5.4 Stress-and-measure (SAM) data of base recombination current variation
with drain bias using forward-biased n+ source/p-body junction as the
minority carrier (electron) emitter. CHE stress at VDS= 8.0 V, VGS= 4.0V
and Vsx= 0.0 V for 120 minutes. Measured at VBE= 0.4 V, VGB= -0.5 V,
and VCB= 0.0 V. (a) Raw data showing decrease at VDX >~3 V due to hole
generated by electron impact, (b) CHE generated excess base current, AIB,
and 3(AIb)/9Vcb versus VCB. (c) Estimated interface trap position from the
n+ drain junction boundary. (IpA/V corresponds to N,T = 109 traps/cm2.)


31
the surface space-charge region of the drain/body junction along the n-channel-
perimeter, (3) IB_DE-b2 *n the drain/body junction bulk space-charge region, (4) IB_DE.bl
in the drain/body junction bulk-quasi-neutral region, (5) IB_DE-b2 same as (1) but
outside of the n-channel-perimeter, and (6) Ib-de-cI same as (2) but outside of the n-
channel-perimeter. Thus, the measured IB.DE is a sum of the 8 components described
above.
^B-DE ^B-el-bulk ^B-el-contact
^B-DE-cl + IB-DE-s2 + +IB-DE-b2 + ^B-DE-bl +
B-DE^+ ^B-DE-cl (3-1)
Of all the base-current components described in (3.1), only IB_DE.cl an<^ ^b-de-s2 can he
modulated by the gate voltage [59] and are sensitive to oxide charge Q0x built-up and
interface traps NIT generation during hot-carrier stress. Thus the 6 gate-voltage
independent components can be subtracted. The above systematic classification of the
IB components during DE-DCIV measurement also applies for SE-DCIV.
Figure 3.3 shows the components of IB in the BE-DCIV or Bottom-Emitter/Drain-
Source-Collector DCIV measurement configuration for a nMOST fabricated with a n+/p
junction well (in most present production chips, only the pMOST has the p+/n isolation
well junction). During BE-DCIV measurement, the collector-base junction is either
grounded or weakly reverse-biased. Thus, the electron-hole recombination components
along the surface space-charge-region and in the bulk space-charge region of the
collectonfoase (drain/body and source/body) junctions are negligible due to low electron
and hole concentrations in these regions. In fact, thermal generation of electron-hole


5
AVGt and AS) with the stress-generated surface recombination current in submicron
nMOST.
Chapter 7 presents a new interface trap generation model recently formulated by
Sah [42], This interface generation model will be used to analyze the current-
accelerated channel hot carriers results presented in Chapter 8. Analysis will show that
the Si-H bonds at the Si02/Si interface are broken by hot holes in n-channel MOST
during channel hot electron stress and in n/p/n BJT during emitter-base reverse-bias
stress, and by hot electrons in p-channel MOST during channel hot hole stress. The
broken Si-H bond forms the dangling silicon bond which is the electrically active
interface trap.
Chapter 9 delineates four fundamental interband Auger and impact generation
pathways in silicon-based MOST where a 4.25 eV hot hole, initiated by primary
energetic electron, can be injected over the Si02/Si barrier into the gate oxide to cause
positive oxide charging in Si02. Three distinct electron threshold kinetic energies are
found using energy conservation law: (1) 5.37 eV for hot-electron/hot-hole interband
impact generation, (2) 4.25 eV for both hot-electron/thermal-hole impact collision
pathway and hot-electron/thermal-hole exchange recombination pathway, and (3) 3.13
eV for hot-electron-thermal hole Auger recombination pathway. Experiments are
described which verified the existence of the 5.37 eV threshold energy for interband
impact generation pathway.
Chapter 10 presents for the first time an interface trap reduction process, observed
using DCIV measurement, in p-channel MOST during conventional channel-hot-hole


100
50 at Vbe=0.74 V. IcH-electron was obtained by subtracting the bulk and exterior-
perimeter electron currents from the total drain terminal current, ID, measured at four
combinations of emitter and gate voltages: forward and zero bias the substrate-body
junction (with and without the injected electrons from the bottom emitter), and
inversion and accumulation gate voltages (with and without the channel electrons at the
Si02/Si interface).
8.3 Results and Discussions
8.3.1 Long Channel CACHC
Production nMOSTs and pMOSTs supplied by several SRC companies are used to
demonstrate the proposed current-accelerated channel hot electron (CACHE) and
channel hot hole (CACHH) stress methodologies. The nMOST on p-well/n+ buried-
layer has 16nm gate oxide, W/L= 100 (i.m/100 |im, and NAA_well= 1.4xl016 Boron/cm3.
The pMOST on n-well/p-substrate has 14nm gate oxide, W/L= 20 |im/20 fim, and
NDD-weii= 3.9xl016 Phosphorus/cm3. In the SAM experiments, the degradation was
monitored by the Drain-Emitter DCIV (DE-DCIV) method which measures the post
stress base or body current versus gate-base voltage characteristics, IB-VGB, with the
drain/body junction forward biased.
Figure 8.3 shows the measured TTFs (defined below) versus applied drain-source
stress voltage VDS both without (VBE=VBS= 0.0 V) and with (VBE>0 or VBS>0) current
acceleration. The post-stress DCIV curves are shown in the inset of Figure 8.3(a) for
nMOST and Figure 8.3(b) for pMOST. The measured TTF is defined as the stress time
required to give le-peak^B-initiar^ or a 100% increase in IB for nMOST and


24
Body Source
i n
\ \ n+7
P-well
N-epitaxial layer
N+ substrate
K230
HP3488
K486
Gate
I
I
|-[hP3478]
V J
> I
' rf'V-t-f K486
L v J
\ * /
o Substrate
HP3478
T
HP6106
K230
J
Figure 2.8 SAM schematic setup for current-accelerated CHE stress using the
forward-biased source/body junction.


59
components cannot be accomplished from one single IG-VGB measurement alone unless
additional properties of the interface traps are known or assumed. However, this can be
easily separated using the DCIV technique. On the IB-VGB plot shown in Figure 4.4(a),
the post-stressed IB peak increases by about 700 times which translate to about 1012
cm'2 Nit after 5xl018 electron/cm'2 SHE stress. On the Ic^GB shown in the Figure
4.4(b), the post-stressed Ic curve shifted by about 0.35V which translates to a negatively
charged Q0T density of 4.3x1011 cm'2.
Additional examples are given in Figure 4.5(a)-(c) for the bottom-Emitter
configuration which use Ic to monitor negative, positive, and turn-around AQ0T induced
by stress. Figure 4.5(a) is identical to Figure 4.4(b) showing positive AVGB caused by
the negative oxide trap -QOT. Figure 4.5(b) shows negative AVGB (curve 2) from
positive oxide trap +Q0T. The generation of positive oxide trap +Q0t anticipated
from the electron-impact emission of electrons trapped by the neutral oxygen vacancy
[72] for a SHEi condition with VGB= 12V and VDB=VSB= 10V.
V0 + e* -> V0+ + 2e~ (4.17)
Figure 4.5(c) illustrates the QOT turn-around effect as coined by Young [75]. A short
CHEi stress (~lsec) at VGB=VDB= 15V with the source floating injects electron into the
oxide along the entire length of the strongly inverted n-channel because VGB VTH.
The electrons are captured by the existing oxide trap resulting in -QOT (Curve 1). After
an additional 500 sec stress, there is a built-up of positive oxide traps from the impact-
emission pathway described by eqn.(4.17) which compensates the -Q0t and shifts the
curve to the right (Curve 2).


4
accelerate the degradation rate of the BJTs. We will demonstrate the extension of this
current-accelerated BJT stress methodology to MOST under channel hot carriers (CHC)
stress. For MOST, the density of the hot carriers is increased by forward-biasing the
substrate/body or source/body n/p junction during CHC stress This approach will be
known as current-accelerated channel hot carriers stress (CACHC). Results will be
analyzed using a physic-based degradation model.
The methodology adopted in this work will be based on Stress-And-Measure
(SAM) experiments "Stress" refers to conventional substrate hot carrier (SHC) and
CHC stresses or the new CACHE stress proposed and demonstrated in the following
chapter. "Measure" refers to electrical characterization of the device degradation. This
SAM methodology will be described in Chapter 2 where an overview of substrate and
channel hot carriers stresses is presented. Chapter 3 presents a novel direct-current
current-voltage (DCIV) technique, proposed by Neugroschel and Sah [41], to
characterize hot-carrier-induced degradation. This technique employs the surface
recombination current of one of the forward-biased and gate-controlled p/n junctions of
the BJT and MOST structures as the degradation monitor. This measurement
configuration is that of a short-circuited-collector BJT and the base current, IB, is
measured as a function of the gate-base dc voltage. This DCIV technique was applied
to separate the stress-generated oxide and interface traps in submicron MOST which
will be described in Chapter 4. Chapter 5 extends the DCIV technique to spatially
profile the stress-generated interface trap distribution in MOST. Chapter 6 presents an
experimental correlation of the conventional measured electrical parameters (AIDsat,


-AIb/(1A) AIb/(1A)
142
10"6 -
DE
tstress=170kS
i 1 r
-AIb
Vgb=-0.5V
VGB=+0.5V
109 U Vgr=+2.0V
109 I-
10
-12
0.2
0.4 0.6
VEB/(1V)
: se 1
^stress= 170kS
^ 1
t
9
8 _
Vgb=-0.5V
B
8
- o Vgb=+0.5V
M

* Vgb=+2.0V

V

g7
r
-
JX' r
961224MH ^
w
-
j I
(bT
0.8
Figure 10.2 The forward current-voltage characteristics (AIB vs VEB) of the drain and
source junction of the 20/0.7 aspect-ratio pMOSTs in the (a) DE and (b) SE
measurement configurations at three gate biases (VGB= 2.0 V, 0.5 V and
-0.5 V) after the CHH-stressed for 170ks. The two broken lines give the
theoretical reciprocal slopes, n=l and n=2.


98
VSB=O.OV
or -0.7V
SiO?
Gate
Vgs=0.9V
Drain
V0S=4.5V
I to 7.0V
>D
Base
itr^
I Ground
S8SS8S
SiOz Ni-r Si**H
1 A A A ,A.. A A A. A A 1
¡8S8S8S
A
n+
-
CH-source
CH-be ^CH-electron
n+
A
p-Base-well or p-Body
'JD-
Bulk
D-Bulk
A
&
960222CTS
n+Bottom-Emitter
Emitter ^4
VEB=0V or -0.63V or -0.74V
Figure 8.1 Cross-sectional view of a nMOST biased simultaneously in the two current-
accelerated CHE stress configurations. Dots are conduction-band electrons.
Circled dots are valence-band electrons. Circles are valence-band holes
generated by energetic electron impact in the n+ drain [19]. Hollow
triangles are interface electronic traps. Solid triangles are weak interface
bonds which become electronic interface traps when ruptured by hole
capture marked by which releases the bonded hydrogen. Solid arrow with
a dot on its tail designates an electron injected from the forward-biased
bottom emitter or source junctions. Hollow arrow with a circle on its tail
designates the hole pathway. Regular channel electrons and thermal
electrons are not shown. Biases are those used to give the data in Figs.2-3.


97
current density (the current-acceleration method). In the current-acceleration method,
the hot carrier density can be increased by forwarding-biasing a p/n junction while the
hot carrier kinetic energy is controlled by the applied voltages during stress.
Under the traditional channel hot electron (CHE) stress of a n-channel MOST
(nMOST), the high drain and low gate voltages (VD>VG) accelerate the source-injected
channel electrons, denoted by IcH-source *n Figure 8.1, to high kinetic energies (=qlVD -
(vG- VGT)I in the space-charge-region of the reverse-biased drain junction. As shown
in Figure 8.1, these hot electrons can generate interface traps by breaking the Si:H and
Si:OH bonds [42] (the star-filled triangle in Figure 8.1) to release the hydrogen and
leave the dangling silicon bonds which are electron and hole interface traps. The
channel hot electrons will also generate additional (secondary) electron-hole pairs
(KE>Eg = 1.12 eV as shown in [109]) near the drain junction. The secondary hot holes
can break additional Si:H and SiO:H bonds to give more interface traps [42], If the
secondary hot holes have KE > Xh= 4.25 eV (the Si02/Si hole barrier), they can also be
injected into the oxide to positively charge the existing oxide hole traps.
Figure 8.1 illustrates the current acceleration method. It shows that the channel hot
electron current, ICH.eiectron = fcH-source + lCH-be> can be increased in two ways: (1)
increasing IcH-source by forward-biasing the source-body or Source-Emitter (SE) n+/p
junction junction (VBS= 0.7 V), and (2) increasing IGH.be by forward-biasing the
substrate-body or Bottom-Emitter (BE) n+/p junction (VBE= 0.63 V or 0.74 V). Figure
8.2 demonstrates increased channel hot electron current from forward-biasing the
bottom-emitter junction (IcH-eiectron = IcH-be) by a factor of about 12 at VBE=0.63 V and


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a thesis for the degree of Doctor of Philosophy.
Chih-Tang Sah, Chanman
Robert C. Pittman Eminent Scholar and
Graduate Research Professor of
Electrical and Computer Engineering
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a thesis for the degree of Doctor of Philosophy.
Amost Neugroschel
Professor of Electrical and Computer
Engineering
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a thesis for the degree of Doctor of Philosophy.
Toshikazu Nishida
Associate Professor of Electrical and
Computer Engineering
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a thesis for the degree of Doctor of Philosophy.
cm> *
Peter Zory fy yy
Professor of Electrical amHjomputer
Engineering


CURRENT-ACCELERATED
CHANNEL HOT-CARRIER STRESS IN
SILICON METAL-OXIDE-SEMICONDUCTOR TRANSISTORS
BY
KIM-KWONG MICHAEL HAN
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1997


10
Figure 2.2
Flow chart of the SAM control program.


CHAPTER 6
LINEAR REDUCTION OF DRAIN CURRENT WITH
INCREASING INTERFACE RECOMBINATION IN NMOS
TRANSISTORS STRESSED BY CHANNEL HOT ELECTRON
6.1 Introduction
Recently, Neugroschel and Sah proposed a simple and highly sensitive method, the
Direct-Current Current-Voltage (DCIV) method, to monitor the degradation of silicon
MOS transistors (MOSTs) under channel and substrate hot carrier stresses [41]. The
DCIV method uses the gate-voltage-controlled interfacial recombination current IB
(base or body current) to monitor the generation rate of interface traps and the charging
rate of oxide traps in MOSTs. However, the traditional transistor reliability
measurements monitor the changes of the saturation drain current (AID_sat), threshold
gate voltage (AVGT) and subthreshold voltage swing (AS) [33-37, 51-53, 98-99]. In this
chapter, we demonstrate the anticipated linear dependencies of these traditional
characteristics on the DCIV base-body recombination current, i.e. AIDsat, AVGT, AS
AIb.
6.2 Experiments
The n-channel MOSTs (nMOSTs) used have a drawn channel width/length aspect
ratio of LAV = 17.5p.m/0.35|lm, a 60A gate oxide, and separate source, drain, gate, body
contact pads. During the Stress-and-Measure (SAM) experiment runs, the nMOSTs
were stressed by channel hot electrons (CHEs) in the forward configuration (i.e.
75


20
the barrier. This linear CHE stress mode was first employed by Abbas and Dockerty in
1974 to investigate the effect of channel hot electron on the electrical characteristics of
the n-channel MOST [49]. Their investigation showed that electrons can be injected
into the gate oxide and subsequently captured by the oxide traps by operating the
nMOST at the drain current saturation point VD=VG. This bias configuration localized
electron trapping near the drain junction and was attributed as the cause of the positive
gate shift of the drain current-gate voltage characteristics with increasing injection time.
Subsequent experimental investigations and numerical simulations by other researchers
have widely supported this localized electron-trapping model under saturation CHE
stress [50-52, 55]. Typically, MOST fabricated with poor quality gate oxide (high
concentration of electron traps) and with appreciable gate current (high electron
injection efficiency) will degrade faster under saturation-point CHE stress [53].
When VDS>VGS, the transistor is biased deep into the drain-current saturation
range. Unlike the saturation-point CHE stress, the channel electrons are mainly
accelerated by the large electric field in the drain space-charge layer due to VD VDsat
instead of by the small channel electric field. If the kinetic energy of the hot electrons is
greater than ~EG.Si (i.e. VD VDsat > IEG_Si/ql, interband impact generation of electron-
hole pairs occurs. Most of the holes flow into the p-well, giving rise to a body (or well)
current in the p-well terminal (Ix). The observed Ix typically peaks at VDS=2VGS [33-
35, 54-55]. A peak well current will correspond to maximizing the combination of
interband impact generation rate of electron-hole pairs and the efficient collection of
holes by the p-well. This well.current can overload the substrate-bias generator during


into the channel region and passivates the fabrication-residual interface traps over the
channel and source interface.


2
drain electric field and power density constant after scaling. However, the present
transistor scaling trend has followed the constant-voltage scaling law more closely in
order to be compatible with power supply voltage and noise margin constraints. But,
scaling down the gate oxide thickness and channel length at constant supply voltage
unavoidedly increases the oxide and channel electric fields. These high fields reduce
MOST reliability because they cause (1) acceleration of hot carriers (electrons or holes)
to high kinetic energies, (2) injection of hot carriers via surmounting the Si02/Si energy
barrier into the gate oxide, (3) quantum-mechanical band-to-band and band-trap-band
tunneling of thermal carriers into the gate oxide, (4) charging and generation of oxide
traps by the injected carriers, and (5) generation of interface traps [15-28]. The physical
charging-discharging and generation-annihilation of the oxide and interface traps by the
hot carriers have been identified to have detrimental effects on the stability and
reliability of the silicon MOSTs. These hot carriers effects cause a gradual degradation
of the transistor electrical characteristics as reflected by changes in transistors threshold
voltage (AVgt), drain-current driving capability (AIDsat), subthreshold distortion (AS)
and an increase in standby power consumption. When the degradation of the MOST
exceeds an acceptable level in a VLSI/ULSI circuit, for example AVGT=10 mV or
^Dsat^DsatO=-^ the circuit may not operate properly and reliable operation of the
electronic system using this circuit will be adversely affected. Therefore, the hot carrier
effects in deep submicron MOST must be understood and minimized such that the
scaled-down MOSTs can continue to achieve a reliable operation Time-to-Failure (TTF)
of 10 years. Innovative hot-carrier-resistant transistor designs and careful processing


CHAPTER 1
INTRODUCTION
Recent advancement in silicon Very-Large-Scale-Integration (VLSI) and Ultra-
Large-Scale-Integration (ULSI) technology, particularly lithography, has led to an
aggressive scaling of metal-oxide-semiconductor transistors (MOSTs) to smaller
dimensions to increase the packing density (transistors/chip) and to integrate more
complex circuit functions onto a single silicon chip or die [1]. The small transistor size
also greatly improves the speed of the VLSI/ULSI circuits. For instance, in the 1980s,
a typical channel length in the MOST VLSI circuit is of the order of lfim [2-3].
Currently, MOSTs fabricated with 0.25(Jxn to 0.35itm channel length are being used to
manufacture high density 64Mb DRAM and fast 200 MHz microprocessor with 107-108
MOSTs [4-6]. Recently, MOSTs with 0.1|im channel length fabricated at research
laboratories have been reported [7-10].
Constant-field and constant-voltage scalings are the two limiting approaches used
for scaling the MOST to increase transistor density and to improve circuit performance
[8, 11-14], Constant-field scaling, first introduced by Dennard of IBM in 1974 [11],
proposed a linear reduction of all the physical geometries of the transistors (oxide
thickness, channel length, channel width, and junction depth) and supply voltage with
increased doping concentration to scale MOST to smaller features. This approach
attempts to maintain a long-channel characteristic in scaled MOST and also to keep the
1


64
Collector
Emitter
Figure 5.1 (a) Cross-sectional view of the nMOST transistor used in the SE-DCIV
profiling configuration, (b) Base recombination current versus drain bias,
Ix VDX or Ib VCB, for five interface trap distributions. Curve-#(Drain,
Channel, Source) are: Curve-0 (0,0,0), Curve-1 (nonuniform, 0, 0), Curve-2
(0, constant, 0), Curve-3 (0, nonuniform, 0), and Curve-4 (nonuniform,
nonuniform, 0).


103
^B-peakAB-initiai=0-2 or a 20% increase in IB for pMOST. The growth of the IB.peak with
increasing stress time is attributed to interface trap generation [41, 96, 101, 111] by hot
carriers (hot holes for nMOST and hot electrons for pMOST) breaking the Si:H and
SiO:H bonds near the drain junction [42], The profiles of the IB peak at VGB= -0.4 V
for nMOST and VGB= 0.35 V for pMOST remain nearly-constant indicating that
negligible interface traps were being generated over the channel. The VGS-shift of the
¡B-peak attributed to charging of the oxide hole traps (negative-VG shift in nMOST)
and electron traps (positive-VG shift in pMOST) from barrier-surmounting injection into
the oxide of the secondary hot carriers interband-impact-generated by the (primary)
channel hot carriers which gave the predicted threshold, 4.25 eV for holes and 3.12 eV
for electrons. During the CACHC stress using the forward-biased bottom emitter
junction, a constant increase in the baseline of the IB current is observed. This is
attributed to the interface trap generated along the exterior (to the channel) peripheral of
drain junction underneath the field oxide. This excess IB component created during
CACHC stress are subtracted in order to account for the actual CHE-induced
degradation occurring at the drain junction interior to the channel.
Figure 8.3(a) contains three sets of TTF data: (1) the traditional CHE stress
without current-acceleration (CA) (four circles for VDS_stress= 5.5 V to 7.0 V); (2)
CACHE stress from forward-biasing the bottom emitter substrate/body n+/p junction
(five solid circles at VBE= 0.63 V for VDS_stress= 5.0 V to 7.0 V giving about 16x
acceleration, and one solid triangle at VBE= 0.74 V for VDS= 4.5 V giving 50x
acceleration); and (3) CACHE stress from forward-biasing the top-emitter source/body


89
subsequent section. Eqns.(7.11) and (7.12) apply only to the drain-current range of the
MOST where VR=IVD VDsatl > 0 where VDsat is the drain voltage at drain current
saturation and may approximated by IVG VGTI. The VBI must also be modified to take
into account of the built-in potential barrier height difference between the drain/body
and source/body junctions. For identical drain and source junction, VBI=0 when the
MOSTs channel is inverted.
Another possible pathway for hole to generate interface traps is via the Interband
Tunneling Hole Injection (labeled as ITHi in Figure 7.1). This ITHi pathway will
become important in heavily doped n++emitter/p+base junction of advanced submicron
BJTs and future generation of MOSTs where the body doping concentration is
approaching and exceeding 1018 cm-3. At this high concentration, valence electrons in
the body(base) region can quantum mechanically tunnel into the empty states of the
conduction band in the quasi-neutral drain (emitter) region, leaving behind thermal
holes. Similarly to the interband impact-generated holes, these interband tunneling
holes can then accelerate in the high-field SCL to create new interface traps. The bond
breaking distance associated with this ITHi interface trap generation pathway is given
as:
-BB
\ (2£/qNAA) [i VR+VBI Vsi_Gap VFN+
-I
VR + VBI VSi-Gap v FN+
3 ^FN+ ^BB 1 ("7-13)
In Eqn.(7.10), the valence-electron tunneling process decreases the voltage dependence
of the bond breaking distance by (VSi.Gap + VFN+) where VSi_Gap
is the silicon energy


i-12
47
O
CD
3
2
1
O i i i i l i i i i I i i i i I i i i i
-1.0 -0.5 0 0.5 1.0
tiii|iiii|iiii|iiir
Top-Emitter BJT
AIb =(qAS0n/2)exp(qVBE/2kT)=1 pA
- S0 =2cm/sec => N,T=109crrf2
i i i i l i i i i I i i i i I i i i i_
VGB/ (1V)
Figure 3.9 Sensitivity test of the top-emitter IB measurement of interface traps of an
unstressed BiMOS with nMOST and npn BJT at VEB= -0.3 V and 297 K.
XqX= 15 nm. W/L= 100 |im/1.6 (im.


Ill
VS=VG/(1 V)
Figure 8.5 Degradation characteristics of 0.35 |im nMOST under CHE stress. IDsat is
measured in the "reverse" (i.e. actually ISsat was measured) configuration to
monitor the degradation near the drain junction. CHE stress condition:
VDS= 3.65 V, VGS= 0.5 V and VSB= 0.0 V.


13
and saturation ID measurement, the source, body (or p-well) and the substrate of the
MOST are grounded. For the linear ID-VG measurement, a programmable voltage K230
is supplied by K230 to give a low (<0.5V) and constant voltage across the drain and
source, VDS, and another K230 voltage source is connected to the gate to step the gate
voltage, in small incremental step, from accumulation to inversion. The drain current is
measured by a sensitive pico-ammeter (K486). The voltages applied to the gate and
drain terminals are measured by the HP34401 digital voltmeter. Thus, the ID-VG
characteristics of the MOST is obtained by measuring the drain current for a range of
gate voltages.
During the saturation IDsat-VD measurement, the gate and drain are tied together
and connected to a single K230 voltage source (see Figure 2.4). This connection
ensures that the MOST is operating in the saturation mode. The drain current IDsat can
either be measured by the K486 picoammeter or the HP3478 ammeter. This flexibility
is necessitated by the 2mA current compliance limit of the K486 picoammeter. The
voltage supplied by the K230 is measured by the HP34401 voltmeter. By measuring the
drain current for a range of drain voltages, the saturation lDsat-VD characteristics can be
obtained.
2.3.2 DCIV Measurement
The DCIV measurement, acronyned by Sah [41], is a plot of the d.c. body (or base)
current versus gate voltage with a forward-biased p/n junction supplying the minority
carriers to recombine with the majority carriers at recombination centers at the Si02/Si
interface of the MOST. In this section, we will only describe the basic measurement


76
n+drain/p-body junction is reverse-biased) while their electrical characteristics were
measured after each stress duration in the reverse configuration (i.e. n+source/p-body
junction is reverse-biased to measure AIs.sat, AVGT and AS, and the n+drain/p-body is
forward-biased to measure AIB in the DCIV mode which is known as the Drain-Emitter
DCIV or DE-DCIV ). This interchange of source and drain during measurements is
traditionally employed because the generated interface traps and charged oxide traps are
located over the drain-junction space-charge region. Two bias voltage conditions are
used during the CHE stress duration: (i) the traditional VDS = 2VGS which gives the
maximum body current, Ix-Stress> anc* (ii) the constant VDS and VGS with
Vds>>VGs-Vgt to probe a wide range of hot electron kinetic energies. In each SAM
experiment, the drain saturation current is measured at VDS=VGS= 1.0 V and Vsx= 0.0
V. The threshold voltage is defined as the gate voltage at ID = 500 |iA with VDS= 0.5 V
and Vsx= 0.0 V. The subthreshold slope is computed at ID = 100 nA with VDS= 0.5 V
and VSx= 0 0 V. The n+drain/p-body junction is forward biased at 0.4 V and the
n+source/p-body is short-circuited for the IB-VGB measurements in the DE-DCIV mode.
6.3 Results and Discussion
Figure 6.1 shows typical IB-VGB DE-DCIV characteristics with increasing
accumulated stress time of a nMOST stressed by CHEs. The stressed voltages were
VDS= 3.3 V and VGS= 1.6 V, corresponding to peak body current condition. The IB_peak
increases with CHE stress time and occurs at a nearly constant VGB = -0.45 V. The rise
of IB_peak *s due to the stress-generated Si02/Si interface trap over the drain junction
space-charge region [84-88, 96]. The negligible VG shift of the IB indicates


52
Top-Emitter Configuration
Bottom-Emitter Configuration
Ct
OV I
ov
1
G
SiO,
n+Coflector
EB5
Q|T bOT
Ay..ifa.i i1 rnrti
B
ov
p+Basc
Contac
pBase
n-epitaxial layer
n+Emitter
Ie
-0.3V
Figure 4.2 Cross-sectional views of the top-emitter and bottom-emitter DCIV
measurement configurations. The bias used in the experiment are also
depicted in the figure.


40
potential at which the recombination rate peaks is negatively displaced as the surface
minority carrier (electrons) density increases due to the injection of minority carriers
from the forward-biased emitter. The surface potential V|/S.max corresponds to peak
surface recombination rate is [59]
V^s-max = vf vbe^2 (kgT/2q) log (cps/cns) (3.12)
At this surface potential, the surface is intrinsic, i.e., the electron and hole
concentrations are the same and equal to
^BE
Ns = Ps = ni exp( ) (3.13)
2kBT
Assuming Ep E¡ = 0, c0 = cns = cps and Ps = Ns > n¡,
CO^ITni ^^BE
Rss = ( ) exp ( ) (3.14)
2 2kBT
This simplification implies that under maximum recombination condition, IB-be-c1 ^
exp(qVBE/2kBT). Next, consider the case for Ps > Ns > n¡,
P-ss = co ^it Ns (3.15)
ql/rs-VF+VBE
= CoN-pgripexpl ) (3.16)
kBT
Here, IB.BE_cl ~ exp(qVBE/lkBT).
3.5 Relationship between \|/s and VGB Under Non-Equilibrium Condition
So far, we have only considered the relationship of Rss as a function of V|/s. It is
necessary to be able to relate the applied gate voltage (VGB) during DCIV measurement
to the surface band-bending potential V|is. This can be obtained by modifying the
boundary conditions when solving the Poisson Equation, Eqn. (3.17), for Vj/S.


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
ABSTRACT vi
CHAPTERS
1 INTRODUCTION 1
2 STRESS-AND-MEASURE (SAM) METHODOLOGY 7
2.1 Introduction 7
2.2 SAM Setup 7
2.3 Electrical Measurement 9
2.4 Substrate and Channel Hot Electron Stress 15
3 DIRECT-CURRENT C U R R E NT V O L T A G E (DCIV)
CHARACTERIZATION TECHNIQUE 26
3.1 Introduction 26
3.2 Measurement Configurations 27
3.3 Components of IB 29
3.4 Theory of Surface Recombination Rate 33
3.5 Relationship between \|/s and VGB under Non-Equilibrium
Condition 40
3.5 Summary 46
4 SEPARATION OF INTERFACE AND OXIDE TRAPS USING THE
DCIV TECHNIQUE 48
4.1 Introduction 48
4.2 Experiment 49
4.3 Results and Discussions 55
iii


141
junction [59]. At the higher forward bias (0.6 V), AIB is dominated by electron-hole
recombination at the interface traps located over the surface channel region between the
source and the drain [59].
The I-V characteristics of the forward-biased drain-emitter (DE) and source-
emitter (SE) junctions further help to locate the interface traps. The reciprocal slope or
ideality factor, n, in AIB = exp(qVEB/nkT), is read from Figure 10.2 and copied to
Figure 10.1. The value n=2 signifies recombination in the space-charge region, while
n=l, in the quasi-neutral (channel) region [59].
It is evident from the DE-DCIV of Figures 10.1(a) and 10.1(b) that the rapidly
growing peak at VGB= 2.0 V with n=2 originates from CHH-generated interface traps
located over the p+drain/n-well surface space-charge-region. The high growth rate
reflects the high density of primary hot holes and secondary hot electrons in this
reverse-biased space-charge-region. The location of the CHH-generated interface traps
is further confirmed by the SE-DCIV of Figure 10.1(c) at VSB= 0.4 V which shows a
residual-fabrication interface traps over the p+source/n-well surface space-charge-region
(because n=2) that is overwhelmed by the mid-channel recombination (n=l) component
when VSB is increased to 0.6 V in Figure 10.1(d). The shift of the n=2 peak in the
positive VGB direction, as indicated in Figures 10.1(a) and 10.1(b), shows the negative
charging of the oxide traps over the drain-junction space-charge region by barrier-
surmounting (Si02/Si xe= 3.12 eV) secondary hot electrons generated by the primary
or channel hot holes in the p+drain, similar to that experimentally proven for positive
charging of the oxide traps during CHE stress in nMOST presented in Chapter 9.


44
VGB/(1V)
Figure 3.7 Relation between surface band-bending potential (ys) and applied gate
voltage (VGB) under equilibrium (VBE= 0 V) and non-equilibrium (VBE=
0.2, 0.3, 0.4 and 0.5 V) conditions. Parameters used: xox= 15 nm, NAA=
1016 cm"3, Vra= 0 V, and T= 300 K.


62
capability of this technique is confined to region very close to the drain junction.
Furthermore, the sensitivity was limited by very low stress-generated currents (~ 5fA)
superimposed on large unstressed leakage current. Ancona et al [85], Heremans et al
[86], and Ma et al [87-88] employed the charge-pumping current technique [89-92]
which is a variation of the well-known pulse field-effect method [93-95]. The spatial
distribution of QIT is obtained by correlating the drain voltage pulse during the charge
filling (accumulation) cycle with the width of the surface space-charge layer. However,
as the charge-pumping technique is based on a dynamic electron-hole capture and
emission events at the interface traps, interpretation of charge-pumping results critically
depends on the shape, rise and fall times, amplitude, base reference level, and frequency
of the gate pulse and the drain/source biasing condition to control the electron-hole
emission and capture rates [86]. Ancona et al [85] and Ma et al [88] also cautioned the
use of proper pulse alignment between the gate and drain (180 out-of-phase) to avoid
erroneous interpretation of the QIT spatial distribution. In this chapter, we will show
that the DCIV method can be extended to profile spatial distribution of stress-generated
Qrr t96]-
5.2 Theory
The DCIV Profiling Technique employs the IB_1C components to determine the
lateral distribution of the interface traps. The magnitude of IB.,C is the areal integration
of the surface recombination rate Rss multiplied by the electronic charge q. Assuming a
constant channel width W, IB.,C can be expressed as


163
[132] C.-T. Sah, Jack Y.-C. Sun, and Joseph J-T Tzou, "Generation-Annealing
Kinetics and Atomic Models of a Compensating Donor in the Surface Space
Charge Layer of Oxidized Silicon," J. Appl. Phys., vol.54, No.2 pp.944-956,
Feb 1983.
[133] C.-T. Sah, Jack Y.-C. Sun, and Joseph J-T Tzou, "Deactivation of the Boron
Acceptor in Silicon by Hydrogen," Appl. Phys. Lett., vol.43, no.2, pp.204-
206, Jul 1983.
[134] C.-T. Sah, Samuel C.S. Pan, and Charles C.H. Hsu, "Hydrogenation and
Annealing Kinetics of Group-Ill Acceptors in Oxidized Silicon," J. Appl.
Phys., vol.57, No.12 June 1985.


CHAPTER 3
DIRECT-CURRENT CURRENT-VOLTAGE (DCIV)
CHARACTERIZATION TECHNIQUE
3.1 Introduction
In the MOST with a n/p junction isolation well (usually the pMOST of the present
production CMOS chips), the vertical or lateral BJT action is attained by forward
biasing either the n+/p well junction or the source/well or drain/well p/n junction. In the
MOST without the n/p junction isolation well (usually the nMOST of the present
production CMOS chips), only the lateral BJT action can be attained by forward-biasing
the drain/body or source/body p/n junction.
In this work, a simple and high sensitivity Direct-Current Current-Voltage (DCIV)
technique is employed to characterize the hot-carrier-induced degradation in silicon
metal-oxide-semiconductor transistors (MOSTs). In this chapter, the basic
measurement technique is discussed. This DCIV technique is based on the gate voltage-
controlled electron-hole recombination rate at the Si02/Si interface traps. This
technique was first employed by Sah in 1961 to study the surface recombination and
channel on silicon transistors [59-60]. Recently, this technique has been re-introduced
by Sah and Neugroschel to study the effect of the stress-created oxide (NOT) and
interface (NIT) traps on the reliability of advanced MOS and Bipolar junction transistors
(MOSTs and BJTs) [39-41,61-63, 96],
26


00/9IV (V6-OU)/ aIV
140
10
i i i i | i i i i | i i i i | i i i I | I I 1 I
. SE:Vsb=0.4V
o 500 sec I 1000 sec
. 500 sec I *
o 100 sec
10 sec X
2000 sec
10 i i i i | i i i i | i i i i
D&VDb=0.6V n_2
2k
' 1 | T 1 1 1
10
I I 1 1 | 1 1 1 1 | I i 1 1 | 1 1 1 1 | 1 1 1 1
. SE:Vsb=0.6V
7 2000 sec
1000 sec
500 sec
5
10 sec _
o 100 sec
o 100 sec
10 sec
a 500 sec
o 1000 sec
n=1 2000 sec n1
0
10 iirii-nmnrniiiinmniiiipriifi-|]'niif
(b)
in
l
2k (d)"
1 i 1 i i 1 i i 1
-10 12 3
VGB/(1V)
4-10 1 2 3
Vqb/(1V)
Figure 10.1 The 400mV and 600mV Drain Emitter (DE) and Source Emitter (SE)
DCIV characteristics of silicon pMOSTs with W/L=20|im/0.7p.m showing
the VGB dependence of AIB at VCB=0 and two measurement forward biases
(Veb=0.4V and 0.6V) after channel hot hole stressed at VDS=-5.25V,
VGS=-1.0V, and Vsx=0V for 10, 100, 500, 1000 and 2000 seconds.


70
Figures 5.3(a) and 5.3(b) shows the AIB-VBE curves at three different gate biases
(Vgb=-0.5V, -1.0V, and -2.0V) for SE- and DE-Ib-Vbe measurement configuration,
respectively. In Figure 5.3(a), the increase in AIB for VGB=-0.5V is primarily
contributed by the surface recombination along the channel, i.e., IB. 1C component. The
surface band-bending potential ys at VGB=-0.5V is computed to be equal to 0.08V,
which is close to the flat-band condition (VGB=-0.65V) obtained by a high-frequency
Cg-VG measurement on a 450|imx450fim capacitor. At low VBE bias, it is observed the
ideality factor n is about 1.0. This suggests that the hole surface concentration Ps is
greater than the surface electron concentration Ns which results in the surface
recombination rate to be dependent on VBE. Thus AIB c IB 1C oc exp(qVBE/kBT). At
high VBE bias, it is observed that there is a change of the ideality factor from n=l to
n=2. This is consistent with Rss exp(qVBE/2kBT) for Ps = Ns, as discussed in
Chapter 3.2. The transition from n=l to n=2 occurs at about 0.55V and agrees
reasonably well with the theoretical value predicted using a simple single energy-level
interface trap model.
VBE = 2 [VF yrs (VGB=0.5V) ] (5.9)
= 2 [ (kBT / q) loge (N^/) 0.08V]
= 2[0.36V 0.08V]
= 0.56V
In Figure 5.3(b), the increase in AIB for three gate bias values is attributed to the surface
recombination within the drain/body surface space-charge-layer. This is supported by


71
Figure 5.3 Effect of VGB on the AIB-VBE characteristic for (a) Source-Emitter BJT and
(b) for Drain-Emitter BJT. All measured at VCB= 0 V and 297K.


36
Substituting Eqns.(3.5) and (3.6) into Eq.(3.4) and using n,p, = nj2, Rss can be
simplified as
(NSPS rii2)
Rss Nit (3.7)
cps (Ns + nx) + cns (Ps + px)
The surface recombination current IBS is then the areal integration of the surface
recombination rate Rss along the y- and z-direction.
lbs
Rss dy dz
(3.8)
Equations (3.4) and (3.7) show that there are three factors that will affect the
characteristics of the surface recombination currents. They are: (1) the spatial interface
trap density, NIT(y,z), (2) the spatial surface electron and hole carrier densities, Ns(y,z)
and Ps(y,z), and (3) the electron and hole capture and emission coefficients, cns, cps, ens
and eps, and the energy level of the interface trap in the silicon energy gap, Ej. During
hot carrier stress, an increase in the density of the stress-generated interface traps will
increase the magnitude of the surface recombination currents. This is the basis of the
DCIV which employs the surface recombination component to monitor hot-carrier-
induced degradation in MOSTs. The capture and emissions coefficients are assumed
not to be affected by the hot-carrier stress.
Several salient characteristic properties of the surface recombination rate Rss are
best understood by considering the IB.c! components, shown in Figures 3.2 and 3.3. In
the following, we will use the IB_cl component, measured using the Bottom-Emitter
DCIV configuration, as an example in our discussion. Consider the transition energy


22
between the depletion point and the drain, attracting the impact-generated holes to the
Si02/Si interface and repelling the electrons away from the Si02/Si interface. A
fraction of the hot holes that gain sufficient energy to surmount the Si02/Si hole barrier
may be injected into the gate oxide, but poor hole injection efficiency is expected due to
the larger 4.25eV hole barrier height. Chapter 9 will elaborate more on this channel hot
hole injection process.
Two types of CHE stress configuration are employed here: conventional CHE and
current-accelerated or CACHE. Figure 2.7 illustrates the schematic setup for
conventional CHE stress. During conventional CHE stress, the source and body are
grounded. Two K230 voltage sources are separately connected to the gate and drain.
The drain current is usually greater than 2mA for wide and short MOST. Thus the
HP3478A multimeter has to be used, instead of the K486 picoammter, to measure the
drain current. A K486 connected in series with the gate measures the gate current
during stress.
For current-accelerated or CACHE stress, the source/body or substrate/body
junction is forward-biased to supply additional hot electrons to increase the hot-
electron-induced degradation rate of the MOST. Thus, the drain and gate setups are
identical to conventional CHE stress. Figure 2.8 depicts the schematic setup for
current-accelerated CHE stress using the forward-biased source/body junction. Two
K230 power supplies are connected to the drain and gate terminals and referenced to the
grounded source. A positive voltage, supplied by a HP6106 power supply, is connected
to the body of the n-channel MOST to forward-bias the p-body/n+source junction.


116
1010
109
108
oT 107
C* 106
105
- 104
103
102
101
0.2 0.3 0.4 0.5 0.6
i/vDS/(i v-1)
Figure 8.10 The traditional voltage-accelerated TTF extrapolation plot for the 2.5V
0.35|im nMOST technology.


27
3.2 Measurement Configurations
The DCIV technique is a gate-controlled BJT measurement and employs the lateral
or vertical BJT of the silicon MOST structure. This was once referred to as BiMOST
[41] which will be discarded and replaced by the above general description [63, 71],
Figure 3.1 shows a cross-sectional view of a nMOST illustrating the vertical and lateral
BJTs. Figure 3.1 illustrates three n/p junctions that can be forward-biased the emitter to
inject minority carriers into the p-base well: (1) the Drain/Base (Body) n+/p junction,
(2) the Source/Base (Body) n+/p junction, and (3) the Substrate/Base (body) n+/p
junction. In this work, the body and base are used interchangeably in the discussion
with the same subscript, B or b. Thus, there are three basic DCIV measurement
configurations: two top-emitter (1) Drain-EmittersDE, or (2) Source-Emitter=SE, and
(3) bottom-emittersBE (Substrate-Emitter is not used to avoid confusion with Source-
Emitter abbreviation). In the DE-DCIV configuration, the n+drain/p-body is the
forward-biased emitter junction and the n+source/p-body is the short-circuited
n+collector/p-base junction. Similarly, in the SE-DCIV configuration, the
n+source/p-body junction is the forward-biased emitter junction and the
n+drain/p-body is the short-circuited n+collector/p-base junction. In the BE-DCIV
configuration, the n+substrate/p-body is the forward-biased emitter junction and the
n+drain/p-body and/or n+source/p-body is/are the short-circuited n+collector/p-base
junction(s). This systematic classification, acronyned by Sah [63, 71], of the DCIV
measurement configurations is necessary to exemplify the physical location of the
CHC-induced degradation, as will be discussed later.


CHAPTER 7
INTERFACE TRAPS GENERATION MODEL
7.1 Introduction
As shown in Chapters 5 and 6, interface traps are generated when MOS transistors
are stressed under channel hot carrier stress conditions. These interface traps are
electron-hole recombination-generation sites and thought to be dangling silicon and
oxygen bonds (Si., SiO.) [101] created by energetic electrons and holes breaking the
weak intrinsic (Si:Si, SiO:Si) or impurity (Si:H, SiO:H) bonds. However, the physical
mechanism of interface trap generation during channel hot carrier stress was still not
well understood. No agreement currently exists on this point in the literature. Some
studies attributed interface trap generation to electron injection [35, 102-103] where a
critical energy of 3.5 eV is needed for the electrons to be injected into the oxide to break
the Si:H bond [35]. Others propose that hot holes are the responsible carriers [104-106],
There was also some independent evidence that recombination of electrons with trapped
holes in the oxide also generate interface traps [7-8]. It was also proposed that the
energy released by electron-hole recombination in the Si02 dissociates a weak Si:H
bond, thus the generation of interface traps involve hot electrons, hot holes, and
hydrogen [106].
In this chapter, a new interface trap generation model, formulated by Sah [42], is
presented to account for the channel-hot-carrier-induced degradation. This model is
83


37
band diagram of a nMOST shown in Figure 3.5 biased in the bottom-emitter DCIV
configuration, the gate voltage modulates the amount of surface energy band bending
(\j/s) at the Si02/Si interface and the forward-biased bottom emitter injects minority
carriers (electron) into the p-base well. The rate of electron-hole recombination at the
interface traps, Rss, will depend on the surface electron and hole concentrations, Ns and
Ps. As indicated in Figure 3.5, they are given by
N
s
and
N.
ni
exp
exp
exp
exp
q(FN Fp)
kBT
kBT
fq(ys
- vF
+ FN FP)
kBT
q(^s VF + VBE )
kBT
(3.9)
(3.10)
(3.11)
-qfs
Ps = PE exp ( ) (3.12)
kBT
In Eqns. (3.9) and (3.11), NE and PE are the equilibrium electron and hole
concentrations. Low injection level is also assumed, i.e. N and P < Pe=Pe=
niexp(q\|/F/kBT).
Figure 3.6 illustrates the relation between the surface recombination rate and
surface potential at the Si02/Si interface by plotting Rss versus \|/s for four different
forward-bias voltages. The shape of the Rss curves is the same but the magnitude of
Rss depends exponentially on the excess surface carrier (electron) density. The surface


58
stress Ic-VGB curve only exhibits a lateral and almost parallel shift from the pre-stress
curve with nearly identical height for both the top-Emitter and bottom-Emitter
measurement configurations. This lateral positive gate voltage shift AVGB is attributed
to the charging of the neutral oxygen vacancy centers as shown in eqn.(4.3). Thus the
VGB-ot components in Eqn.(4.1) can be determined from
aqxt/ q
(4.11)
(C0 / q) AVqB_0T
(4.12)
(C0/q)[-0.2 (-0.55)]
(4.13)
4.3xl01;Lcm'2
(4.14)
The effect of hot-electron-induced degradation on the nMOSTs characteristics
can also be observed from the conventional Id-Vgb measurement shown in Figure
4.4(c). After injecting the gate oxide with an electron fluence of 5xl018cm2, the post-
stressed ID current decreases with increasing VGB and there is clearly a distortion in the
subthreshold regime. The subthreshold slope change AS can be employed to estimate
the density of interface traps using the following well-known equation [73].
Dit = (C0/q) (q/2.303kBT) AS (4.15)
= 1012cm_2eV-1 (4.16)
However, the ID degradation is clearly due to the stress-generated AQIT and AQ0T
whose presence is unambigously confirmed by the IB-VGB and IC-VGB measurement.
This is depicted in Figure 4.4(c) where the amount of total gate voltage shift AVGB
extracted at a constant drain current is shown to consists of two components AVGB_0T
and AVgb_it. The separation of the total gate voltage shift AVGB into these two


32
Body
Base
V
GB
Gate
Drain
Collector
lc
Nogjlec NoT.hoij gjQ
'2 Nit N|T:H
A ^aaaaaI
SiO?
p-Body
p-Well
i mriTnTrrm
IShs n+Substrate
(1)IB
x>
mmtt
n+
35?
=J>0=
1X^^=
(3)lB^ffg=
(6)1
TFRT
s
V^-mvI ^bstrate
BE Emitter
Figure 3.3 Cross-sectional view of n-channel MOST biased in the Bottom-Emitter
(BE) DCIV configuration. Three base current recombination pathways are
depicted. See Figure 3.2 caption for the description of symbols.